Misuse and misplaced trust of AI is becoming an unfortunate common event. For example, lawyers trying to leverage the power of generative AI for research submit court filings citing multiple compelling legal precedents. The problem? The AI had confidently, eloquently, and completely fabricated the cases cited. The resulting sanctions and public embarrassment can become a viral cautionary tale, shared across social media as a stark example of AI’s fallibility.

This goes beyond a technical glitch; it’s a catastrophic failure of trust in AI tools in an industry where accuracy and trust are critical. The trust issue here is twofold — the law firms are submitting briefs in which they have blindly over-trusted the AI tool to return accurate information. The subsequent fallout can lead to a strong distrust in AI tools, to the point where platforms featuring AI might not be considered for use until trust is reestablished.

Issues with trusting AI aren’t limited to the legal field. We are seeing the impact of fictional AI-generated information in critical fields such as healthcare and education. On a more personal scale, many of us have had the experience of asking Siri or Alexa to perform a task, only to have it done incorrectly or not at all, for no apparent reason. I’m guilty of sending more than one out-of-context hands-free text to an unsuspecting contact after Siri mistakenly pulls up a completely different name than the one I’d requested.

With digital products moving to incorporate generative and agentic AI at an increasingly frequent rate, trust has become the invisible user interface. When it works, our interactions are seamless and powerful. When it breaks, the entire experience collapses, with potentially devastating consequences. As UX professionals, we’re on the front lines of a new twist on a common challenge. How do we build products that users can rely on? And how do we even begin to measure something as ephemeral as trust in AI?

Trust isn’t a mystical quality. It is a psychological construct built on predictable factors. I won’t dive deep into academic literature on trust in this article. However, it is important to understand that trust is a concept that can be understood, measured, and designed for. This article will provide a practical guide for UX researchers and designers. We will briefly explore the psychological anatomy of trust, offer concrete methods for measuring it, and provide actionable strategies for designing more trustworthy and ethical AI systems.

The Anatomy of Trust: A Psychological Framework for AI

To build trust, we must first understand its components. Think of trust like a four-legged stool. If any one leg is weak, the whole thing becomes unstable. Based on classic psychological models, we can adapt these “legs” for the AI context.

1. Ability (or Competence)

This is the most straightforward pillar: Does the AI have the skills to perform its function accurately and effectively? If a weather app is consistently wrong, you stop trusting it. If an AI legal assistant creates fictitious cases, it has failed the basic test of ability. This is the functional, foundational layer of trust.

2. Benevolence

This moves from function to intent. Does the user believe the AI is acting in their best interest? A GPS that suggests a toll-free route even if it’s a few minutes longer might be perceived as benevolent. Conversely, an AI that aggressively pushes sponsored products feels self-serving, eroding this sense of benevolence. This is where user fears, such as concerns about job displacement, directly challenge trust—the user starts to believe the AI is not on their side.

3. Integrity

Does AI operate on predictable and ethical principles? This is about transparency, fairness, and honesty. An AI that clearly states how it uses personal data demonstrates integrity. A system that quietly changes its terms of service or uses dark patterns to get users to agree to something violates integrity. An AI job recruiting tool that has subtle yet extremely harmful social biases, existing in the algorithm, violates integrity.

4. Predictability & Reliability

Can the user form a stable and accurate mental model of how the AI will behave? Unpredictability, even if the outcomes are occasionally good, creates anxiety. A user needs to know, roughly, what to expect. An AI that gives a radically different answer to the same question asked twice is unpredictable and, therefore, hard to trust.

The Trust Spectrum: The Goal of a Well-Calibrated Relationship

Our goal as UX professionals shouldn’t be to maximize trust at all costs. An employee who blindly trusts every email they receive is a security risk. Likewise, a user who blindly trusts every AI output can be led into dangerous situations, such as the legal briefs referenced at the beginning of this article. The goal is well-calibrated trust.

Think of it as a spectrum where the upper-mid level is the ideal state for a truly trustworthy product to achieve:

• Active Distrust
  The user believes the AI is incompetent or malicious. They will avoid it or actively work against it.
• Suspicion & Scrutiny
  The user interacts cautiously, constantly verifying the AI’s outputs. This is a common and often healthy state for users of new AI.
• Calibrated Trust (The Ideal State)
  This is the sweet spot. The user has an accurate understanding of the AI’s capabilities—its strengths and, crucially, its weaknesses. They know when to rely on it and when to be skeptical.
• Over-trust & Automation Bias
  The user unquestioningly accepts the AI’s outputs. This is where users follow flawed AI navigation into a field or accept a fictional legal brief as fact.

Our job is to design experiences that guide users away from the dangerous poles of Active Distrust and Over-trust and toward that healthy, realistic middle ground of Calibrated Trust.

“

The Researcher’s Toolkit: How to Measure Trust In AI

Trust feels abstract, but it leaves measurable fingerprints. Academics in the social sciences have done much to define both what trust looks like and how it might be measured. As researchers, we can capture these signals through a mix of qualitative, quantitative, and behavioral methods.

Qualitative Probes: Listening For The Language Of Trust

During interviews and usability tests, go beyond “Was that easy to use?” and listen for the underlying psychology. Here are some questions you can start using tomorrow:

• To measure Ability:
  “Tell me about a time this tool’s performance surprised you, either positively or negatively.”
• To measure Benevolence:
  “Do you feel this system is on your side? What gives you that impression?”
• To measure Integrity:
  “If this AI made a mistake, how would you expect it to handle it? What would be a fair response?”
• To measure Predictability:
  “Before you clicked that button, what did you expect the AI to do? How closely did it match your expectation?”

Investigating Existential Fears (The Job Displacement Scenario)

One of the most potent challenges to an AI’s Benevolence is the fear of job displacement. When a participant expresses this, it is a critical research finding. It requires a specific, ethical probing technique.

Imagine a participant says, “Wow, it does that part of my job pretty well. I guess I should be worried.”

An untrained researcher might get defensive or dismiss the comment. An ethical, trained researcher validates and explores:

“Thank you for sharing that; it’s a vital perspective, and it’s exactly the kind of feedback we need to hear. Can you tell me more about what aspects of this tool make you feel that way? In an ideal world, how would a tool like this work with you to make your job better, not to replace it?”

This approach respects the participant, validates their concern, and reframes the feedback into an actionable insight about designing a collaborative, augmenting tool rather than a replacement. Similarly, your findings should reflect the concern users expressed about replacement. We shouldn’t pretend this fear doesn’t exist, nor should we pretend that every AI feature is being implemented with pure intention. Users know better than that, and we should be prepared to argue on their behalf for how the technology might best co-exist within their roles.

Quantitative Measures: Putting A Number On Confidence

You can quantify trust without needing a data science degree. After a user completes a task with an AI, supplement your standard usability questions with a few simple Likert-scale items:

• “The AI’s suggestion was reliable.” (1-7, Strongly Disagree to Strongly Agree)
• “I am confident in the AI’s output.” (1-7)
• “I understood why the AI made that recommendation.” (1-7)
• “The AI responded in a way that I expected.” (1-7)
• “The AI provided consistent responses over time.” (1-7)

Over time, these metrics can track how trust is changing as your product evolves.

Note: If you want to go beyond these simple questions that I’ve made up, there are numerous scales (measurements) of trust in technology that exist in academic literature. It might be an interesting endeavor to measure some relevant psychographic and demographic characteristics of your users and see how that correlates with trust in AI/your product. Table 1 at the end of the article contains four examples of current scales you might consider using to measure trust. You can decide which is best for your application, or you might pull some of the items from any of the scales if you aren’t looking to publish your findings in an academic journal, yet want to use items that have been subjected to some level of empirical scrutiny.

Behavioral Metrics: Observing What Users Do, Not Just What They Say

People’s true feelings are often revealed in their actions. You can use behaviors that reflect the specific context of use for your product. Here are a few general metrics that might apply to most AI tools that give insight into users’ behavior and the trust they place in your tool.

• Correction Rate
  How often do users manually edit, undo, or ignore the AI’s output? A high correction rate is a powerful signal of low trust in its Ability.
• Verification Behavior
  Do users switch to Google or open another application to double-check the AI’s work? This indicates they don’t trust it as a standalone source of truth. It can also potentially be positive that they are calibrating their trust in the system when they use it up front.
• Disengagement
  Do users turn the AI feature off? Do they stop using it entirely after one bad experience? This is the ultimate behavioral vote of no confidence.

Designing For Trust: From Principles to Pixels

Once you’ve researched and measured trust, you can begin to design for it. This means translating psychological principles into tangible interface elements and user flows.

Designing for Competence and Predictability

• Set Clear Expectations
  Use onboarding, tooltips, and empty states to honestly communicate what the AI is good at and where it might struggle. A simple “I’m still learning about [topic X], so please double-check my answers” can work wonders.
• Show Confidence Levels
  Instead of just giving an answer, have the AI signal its own uncertainty. A weather app that says “70% chance of rain” is more trustworthy than one that just says “It will rain” and is wrong. An AI could say, “I’m 85% confident in this summary,” or highlight sentences it’s less sure about.

The Role of Explainability (XAI) and Transparency

Explainability isn’t about showing users the code. It’s about providing a useful, human-understandable rationale for a decision.

Instead of:
“Here is your recommendation.”

Try:
“Because you frequently read articles about UX research methods, I’m recommending this new piece on measuring trust in AI.”

This addition transforms AI from an opaque oracle to a transparent logical partner.

Many of the popular AI tools (e.g., ChatGPT and Gemini) show the thinking that went into the response they provide to a user. Figure 3 shows the steps Gemini went through to provide me with a non-response when I asked it to help me generate the masterpiece displayed above in Figure 2. While this might be more information than most users care to see, it provides a useful resource for a user to audit how the response came to be, and it has provided me with instructions on how I might proceed to address my task.

Figure 4 shows an example of a scorecard OpenAI makes available as an attempt to increase users’ trust. These scorecards are available for each ChatGPT model and go into the specifics of how the models perform as it relates to key areas such as hallucinations, health-based conversations, and much more. In reading the scorecards closely, you will see that no AI model is perfect in any area. The user must remain in a trust but verify mode to make the relationship between human reality and AI work in a way that avoids potential catastrophe. There should never be blind trust in an LLM.

Designing For Trust Repair (Graceful Error Handling) And Not Knowing an Answer

Your AI will make mistakes.

Trust is not determined by the absence of errors, but by how those errors are handled.

• Acknowledge Errors Humbly.
  When the AI is wrong, it should be able to state that clearly. “My apologies, I misunderstood that request. Could you please rephrase it?” is far better than silence or a nonsensical answer.
• Provide an Easy Path to Correction.
  Make feedback mechanisms (like thumbs up/down or a correction box) obvious. More importantly, show that the feedback is being used. A “Thank you, I’m learning from your correction” can help rebuild trust after a failure. As long as this is true.

Likewise, your AI can’t know everything. You should acknowledge this to your users.

UX practitioners should work with the product team to ensure that honesty about limitations is a core product principle.

“

This can include the following:

• Establish User-Centric Metrics: Instead of only measuring engagement or task completion, UXers can work with product managers to define and track metrics like:
  • Hallucination Rate: The frequency with which the AI provides verifiably false information.
  • Successful Fallback Rate: How often the AI correctly identifies its inability to answer and provides a helpful, honest alternative.
• Prioritize the “I Don’t Know” Experience: UXers should frame the “I don’t know” response not as an error state, but as a critical feature. They must lobby for the engineering and content resources needed to design a high-quality, helpful fallback experience.

UX Writing And Trust

All of these considerations highlight the critical role of UX writing in the development of trustworthy AI. UX writers are the architects of the AI’s voice and tone, ensuring that its communication is clear, honest, and empathetic. They translate complex technical processes into user-friendly explanations, craft helpful error messages, and design conversational flows that build confidence and rapport. Without thoughtful UX writing, even the most technologically advanced AI can feel opaque and untrustworthy.

The words and phrases an AI uses are its primary interface with users. UX writers are uniquely positioned to shape this interaction, ensuring that every tooltip, prompt, and response contributes to a positive and trust-building experience. Their expertise in human-centered language and design is indispensable for creating AI systems that not only perform well but also earn and maintain the trust of their users.

A few key areas for UX writers to focus on when writing for AI include:

• Prioritize Transparency
  Clearly communicate the AI’s capabilities and limitations, especially when it’s still learning or if its responses are generated rather than factual. Use phrases that indicate the AI’s nature, such as “As an AI, I can…” or “This is a generated response.”
• Design for Explainability
  When the AI provides a recommendation, decision, or complex output, strive to explain the reasoning behind it in an understandable way. This builds trust by showing the user how the AI arrived at its conclusion.
• Emphasize User Control
  Empower users by providing clear ways to provide feedback, correct errors, or opt out of certain AI features. This reinforces the idea that the user is in control and the AI is a tool to assist them.

The Ethical Tightrope: The Researcher’s Responsibility

As the people responsible for understanding and advocating for users, we walk an ethical tightrope. Our work comes with profound responsibilities.

The Danger Of “Trustwashing”

We must draw a hard line between designing for calibrated trust and designing to manipulate users into trusting a flawed, biased, or harmful system. For example, if an AI system designed for loan approvals consistently discriminates against certain demographics but presents a user interface that implies fairness and transparency, this would be an instance of trustwashing.

Another example of trustwashing would be if an AI medical diagnostic tool occasionally misdiagnoses conditions, but the user interface makes it seem infallible. To avoid trustwashing, the system should clearly communicate the potential for error and the need for human oversight.

Our goal must be to create genuinely trustworthy systems, not just the perception of trust. Using these principles to lull users into a false sense of security is a betrayal of our professional ethics.

To avoid and prevent trustwashing, researchers and UX teams should:

• Prioritize genuine transparency.
  Clearly communicate the limitations, biases, and uncertainties of AI systems. Don’t overstate capabilities or obscure potential risks.
• Conduct rigorous, independent evaluations.
  Go beyond internal testing and seek external validation of system performance, fairness, and robustness.
• Engage with diverse stakeholders.
  Involve users, ethics experts, and impacted communities in the design, development, and evaluation processes to identify potential harms and build genuine trust.
• Be accountable for outcomes.
  Take responsibility for the societal impact of AI systems, even if unintended. Establish mechanisms for redress and continuous improvement.
• Be accountable for outcomes.
  Establish clear and accessible mechanisms for redress when harm occurs, ensuring that individuals and communities affected by AI decisions have avenues for recourse and compensation.
• Educate the public.
  Help users understand how AI works, its limitations, and what to look for when evaluating AI products.
• Advocate for ethical guidelines and regulations.
  Support the development and implementation of industry standards and policies that promote responsible AI development and prevent deceptive practices.
• Be wary of marketing hype.
  Critically assess claims made about AI systems, especially those that emphasize “trustworthiness” without clear evidence or detailed explanations.
• Publish negative findings.
  Don’t shy away from reporting challenges, failures, or ethical dilemmas encountered during research. Transparency about limitations is crucial for building long-term trust.
• Focus on user empowerment.
  Design systems that give users control, agency, and understanding rather than just passively accepting AI outputs.

The Duty To Advocate

When our research uncovers deep-seated distrust or potential harm — like the fear of job displacement — our job has only just begun. We have an ethical duty to advocate for that user. In my experience directing research teams, I’ve seen that the hardest part of our job is often carrying these uncomfortable truths into rooms where decisions are made. We must champion these findings and advocate for design and strategy shifts that prioritize user well-being, even when it challenges the product roadmap.

I personally try to approach presenting this information as an opportunity for growth and improvement, rather than a negative challenge.

For example, instead of stating “Users don’t trust our AI because they fear job displacement,” I might frame it as “Addressing user concerns about job displacement presents a significant opportunity to build deeper trust and long-term loyalty by demonstrating our commitment to responsible AI development and exploring features that enhance human capabilities rather than replace them.” This reframing can shift the conversation from a defensive posture to a proactive, problem-solving mindset, encouraging collaboration and innovative solutions that ultimately benefit both the user and the business.

It’s no secret that one of the more appealing areas for businesses to use AI is in workforce reduction. In reality, there will be many cases where businesses look to cut 10–20% of a particular job family due to the perceived efficiency gains of AI. However, giving users the opportunity to shape the product may steer it in a direction that makes them feel safer than if they do not provide feedback. We should not attempt to convince users they are wrong if they are distrustful of AI. We should appreciate that they are willing to provide feedback, creating an experience that is informed by the human experts who have long been doing the task being automated.

Conclusion: Building Our Digital Future On A Foundation Of Trust

The rise of AI is not the first major technological shift our field has faced. However, it presents one of the most significant psychological challenges of our current time. Building products that are not just usable but also responsible, humane, and trustworthy is our obligation as UX professionals.

Trust is not a soft metric. It is the fundamental currency of any successful human-technology relationship. By understanding its psychological roots, measuring it with rigor, and designing for it with intent and integrity, we can move from creating “intelligent” products to building a future where users can place their confidence in the tools they use every day. A trust that is earned and deserved.

Table 1: Published Academic Scales Measuring Trust In Automated Systems

Appendix A: Trust-Building Tactics Checklist

To design for calibrated trust, consider implementing the following tactics, organized by the four pillars of trust:

1. Ability (Competence) & Predictability

• ✅ Set Clear Expectations: Use onboarding, tooltips, and empty states to honestly communicate the AI’s strengths and weaknesses.
• ✅ Show Confidence Levels: Display the AI’s uncertainty (e.g., “70% chance,” “85% confident”) or highlight less certain parts of its output.
• ✅ Provide Explainability (XAI): Offer useful, human-understandable rationales for the AI’s decisions or recommendations (e.g., “Because you frequently read X, I’m recommending Y”).
• ✅ Design for Graceful Error Handling:
  • ✅ Acknowledge errors humbly (e.g., “My apologies, I misunderstood that request.”).
  • ✅ Provide easy paths to correction (e. ] g., prominent feedback mechanisms like thumbs up/down).
  • ✅ Show that feedback is being used (e.g., “Thank you, I’m learning from your correction”).
• ✅ Design for “I Don’t Know” Responses:
  • ✅ Acknowledge limitations honestly.
  • ✅ Prioritize a high-quality, helpful fallback experience when the AI cannot answer.
• ✅ Prioritize Transparency: Clearly communicate the AI’s capabilities and limitations, especially if responses are generated.

2. Benevolence

• ✅ Address Existential Fears: When users express concerns (e.g., job displacement), validate their concerns and reframe the feedback into actionable insights about collaborative tools.
• ✅ Prioritize User Well-being: Advocate for design and strategy shifts that prioritize user well-being, even if it challenges the product roadmap.
• ✅ Emphasize User Control: Provide clear ways for users to give feedback, correct errors, or opt out of AI features.

3. Integrity

• ✅ Adhere to Ethical Principles: Ensure the AI operates on predictable, ethical principles, demonstrating fairness and honesty.
• ✅ Prioritize Genuine Transparency: Clearly communicate the limitations, biases, and uncertainties of AI systems; avoid overstating capabilities or obscuring risks.
• ✅ Conduct Rigorous, Independent Evaluations: Seek external validation of system performance, fairness, and robustness to mitigate bias.
• ✅ Engage Diverse Stakeholders: Involve users, ethics experts, and impacted communities in the design and evaluation processes.
• ✅ Be Accountable for Outcomes: Establish clear mechanisms for redress and continuous improvement for societal impacts, even if unintended.
• ✅ Educate the Public: Help users understand how AI works, its limitations, and how to evaluate AI products.
• ✅ Advocate for Ethical Guidelines: Support the development and implementation of industry standards and policies that promote responsible AI.
• ✅ Be Wary of Marketing Hype: Critically assess claims about AI “trustworthiness” and demand verifiable data.
• ✅ Publish Negative Findings: Be transparent about challenges, failures, or ethical dilemmas encountered during research.

4. Predictability & Reliability

• ✅ Set Clear Expectations: Use onboarding, tooltips, and empty states to honestly communicate what the AI is good at and where it might struggle.
• ✅ Show Confidence Levels: Instead of just giving an answer, have the AI signal its own uncertainty.
• ✅ Provide Explainability (XAI) and Transparency: Offer a useful, human-understandable rationale for AI decisions.
• ✅ Design for Graceful Error Handling: Acknowledge errors humbly and provide easy paths to correction.
• ✅ Prioritize the “I Don’t Know” Experience: Frame “I don’t know” as a feature and design a high-quality fallback experience.
• ✅ Prioritize Transparency (UX Writing): Clearly communicate the AI’s capabilities and limitations, especially when it’s still learning or if responses are generated.
• ✅ Design for Explainability (UX Writing): Explain the reasoning behind AI recommendations, decisions, or complex outputs.

(yk)

Climate change is the single biggest health threat to humanity, accelerated by human activities such as the burning of fossil fuels, which generate greenhouse gases that trap the sun’s heat.

The average temperature of the earth’s surface is now 1.2°C warmer than it was in the late 1800’s, and projected to more than double by the end of the century.

The consequences of climate change include intense droughts, water shortages, severe fires, melting polar ice, catastrophic storms, and declining biodiversity.

The Internet Is A Significant Part Of The Problem

Shockingly, the internet is responsible for higher global greenhouse emissions than the aviation industry, and is projected to be responsible for 14% of all global greenhouse gas emissions by 2040.

If the internet were a country, it would be the 4th largest polluter in the world and represents the largest coal-powered machine on the planet.

But how can something digital like the internet produce harmful emissions?

Internet emissions come from powering the infrastructure that drives the internet, such as the vast data centres and data transmission networks that consume huge amounts of electricity.

Internet emissions also come from the global manufacturing, distribution, and usage of the estimated 30.5 billion devices (phones, laptops, etc.) that we use to access the internet.

Unsurprisingly, internet related emissions are increasing, given that 60% of the world’s population spend, on average, 40% of their waking hours online.

We Must Urgently Reduce The Environmental Impact Of The Internet

As responsible digital professionals, we must act quickly to minimise the environmental impact of our work.

It is encouraging to see the UK government encourage action by adding “Minimise environmental impact” to their best practice design principles, but there is still too much talking and not enough corrective action taking place within our industry.

The reality of many tightly constrained, fast-paced, and commercially driven web projects is that minimising environmental impact is far from the agenda.

So how can we make the environment more of a priority and talk about it in ways that stakeholders will listen to?

A eureka moment on a recent web optimisation project gave me an idea.

My Eureka Moment

I led a project to optimise the mobile performance of www.talktofrank.com, a government drug advice website that aims to keep everyone safe from harm.

Mobile performance is critically important for the success of this service to ensure that users with older mobile devices and those using slower network connections can still access the information they need.

Our work to minimise page weights focused on purely technical changes that our developer made following recommendations from tools such as Google Lighthouse that reduced the size of the webpages of a key user journey by up to 80%. This resulted in pages downloading up to 30% faster and the carbon footprint of the journey being reduced by 80%.

We hadn’t set out to reduce the carbon footprint, but seeing these results led to my eureka moment.

I realised that by minimising page weights, you improve performance (which is a win for users and service owners) and also consume less energy (due to needing to transfer and store less data), creating additional benefits for the planet — so everyone wins.

This felt like a breakthrough because business, user, and environmental requirements are often at odds with one another. By focussing on minimising websites to be as simple, lightweight and easy to use as possible you get benefits that extend beyond the triple bottom line of people, planet and profit to include performance and purpose.

So why is ‘minimising’ such a great digital sustainability strategy?

• Profit
  Website providers win because their website becomes more efficient and more likely to meet its intended outcomes, and a lighter site should also lead to lower hosting bills.
• People
  People win because they get to use a website that downloads faster, is quick and easy to use because it’s been intentionally designed to be as simple as possible, enabling them to complete their tasks with the minimum amount of effort and mental energy.
• Performance
  Lightweight webpages download faster so perform better for users, particularly those on older devices and on slower network connections.
• Planet
  The planet wins because the amount of energy (and associated emissions) that is required to deliver the website is reduced.
• Purpose
  We know that we do our best work when we feel a sense of purpose. It is hugely gratifying as a digital professional to know that our work is doing good in the world and contributing to making things better for people and the environment.

In order to prioritise the environment, we need to be able to speak confidently in a language that will resonate with the business and ensure that any investment in time and resources yields the widest range of benefits possible.

So even if you feel that the environment is a very low priority on your projects, focusing on minimising page weights to improve performance (which is generally high on the agenda) presents the perfect trojan horse for an environmental agenda (should you need one).

Doing the right thing isn’t always easy, but we’ve done it before when managing to prioritise issues such as usability, accessibility, and inclusion on digital projects.

Many of the things that make websites easier to use, more accessible, and more effective also help to minimise their environmental impact, so the things you need to do will feel familiar and achievable, so don’t worry about it all being another new thing to learn about!

So this all makes sense in theory, but what’s the master plan to use when putting it into practice?

The Masterplan

The masterplan for creating websites that have minimal environmental impact is to focus on offering the maximum value from the minimum input of energy.

It’s an adaptation of Buckminister Fuller’s ‘Dymaxion’ principle, which is one of his many progressive and groundbreaking sustainability strategies for living and surviving on a planet with finite resources.

Inputs of energy include both the electrical energy that is required to operate websites and also the mental energy that is required to use them.

You can achieve this by minimising websites to their core content, features, and functionality, ensuring that everything can be justified from the perspective of meeting a business or user need. This means that anything that isn’t adding a proportional amount of value to the amount of energy it requires to provide it should be removed.

So that’s the masterplan, but how do you put it into practice?

Decarbonise Your Highest Value User Journeys

I’ve developed a new approach called ‘Decarbonising User Journeys’ that will help you to minimise the environmental impact of your website and maximise its performance.

Note: The approach deliberately focuses on optimising key user journeys and not entire websites to keep things manageable and to make it easier to get started.

The secret here is to start small, demonstrate improvements, and then scale.

The approach consists of five simple steps:

1. Identify your highest value user journey,
2. Benchmark your user journey,
3. Set targets,
4. Decarbonise your user journey,
5. Track and share your progress.

Here’s how it works.

Step 1: Identify Your Highest Value User Journey

Your highest value user journey might be the one that your users value the most, the one that brings you the highest revenue, or the one that is fundamental to the success of your organisation.

You could also focus on a user journey that you know is performing particularly badly and has the potential to deliver significant business and user benefits if improved.

You may have lots of important user journeys, and it’s fine to decarbonise multiple journeys in parallel if you have the resources, but I’d recommend starting with one first to keep things simple.

To bring this to life, let’s consider a hypothetical example of a premiership football club trying to decarbonise its online ticket-buying journey that receives high levels of traffic and is responsible for a significant proportion of its weekly income.

Step 2: Benchmark Your User Journey

Once you’ve selected your user journey, you need to benchmark it in terms of how well it meets user needs, the value it offers your organisation, and its carbon footprint.

It is vital that you understand the job it needs to do and how well it is doing it before you start to decarbonise it. There is no point in removing elements of the journey in an effort to reduce its carbon footprint, for example, if you compromise its ability to meet a key user or business need.

You can benchmark how well your user journey is meeting user needs by conducting user research alongside analysing existing customer feedback. Interviews with business stakeholders will help you to understand the value that your journey is providing the organisation and how well business needs are being met.

You can benchmark the carbon footprint and performance of your user journey using online tools such as Cardamon, Ecograder, Website Carbon Calculator, Google Lighthouse, and Bioscore. Make sure you have your analytics data to hand to help get the most accurate estimate of your footprint.

To use these tools, simply add the URL of each page of your journey, and they will give you a range of information such as page weight, energy rating, and carbon emissions. Google Lighthouse works slightly differently via a browser plugin and generates a really useful and detailed performance report as opposed to giving you a carbon rating.

A great way to bring your benchmarking scores to life is to visualise them in a similar way to how you would present a customer journey map or service blueprint.

This example focuses on just communicating the carbon footprint of the user journey, but you can also add more swimlanes to communicate how well the journey is performing from a user and business perspective, too, adding user pain points, quotes, and business metrics where appropriate.

I’ve found that adding the energy efficiency ratings is really effective because it’s an approach that people recognise from their household appliances. This adds a useful context to just showing the weights (such as grams or kilograms) of CO2, which are generally meaningless to people.

Within my benchmarking reports, I also add a set of benchmarking data for every page within the user journey. This gives your stakeholders a more detailed breakdown and a simple summary alongside a snapshot of the benchmarked page.

Your benchmarking activities will give you a really clear picture of where remedial work is required from an environmental, user, and business point of view.

In our football user journey example, it’s clear that the ‘News’ and ‘Tickets’ pages need some attention to reduce their carbon footprint, so they would be a sensible priority for decarbonising.

Step 3: Set Targets

Use your benchmarking results to help you set targets to aim for, such as a carbon budget, energy efficiency, maximum page weight, and minimum Google Lighthouse performance targets for each individual page, in addition to your existing UX metrics and business KPIs.

There is no right or wrong way to set targets. Choose what you think feels achievable and viable for your business, and you’ll only learn how reasonable and achievable they are when you begin to decarbonise your user journeys.

Setting targets is important because it gives you something to aim for and keeps you focused and accountable. The quantitative nature of this work is great because it gives you the ability to quickly demonstrate the positive impact of your work, making it easier to justify the time and resources you are dedicating to it.

Step 4: Decarbonise Your User Journey

Your objective now is to decarbonise your user journey by minimising page weights, improving your Lighthouse performance rating, and minimising pages so that they meet both user and business needs in the most efficient, simple, and effective way possible.

It’s up to you how you approach this depending on the resources and skills that you have, you can focus on specific pages or addressing a specific problem area such as heavyweight images or videos across the entire user journey.

Here’s a list of activities that will all help to reduce the carbon footprint of your user journey:

• Work through the recommendations in the ‘diagnostics’ section of your Google Lighthouse report to help optimise page performance.
• Switch to a green hosting provider if you are not already using one. Use the Green Web Directory to help you choose one.
• Work through the W3C Web Sustainability Guidelines, implementing the most relevant guidelines to your specific user journey.
• Remove anything that is not adding any user or business value.
• Reduce the amount of information on your webpages to make them easier to read and less overwhelming for people.
• Replace content with a lighter-weight alternative (such as swapping a video for text) if the lighter-weight alternative provides the same value.
• Optimise assets such as photos, videos, and code to reduce file sizes.
• Remove any barriers to accessing your website and any distractions that are getting in the way.
• Re-use familiar components and design patterns to make your websites quicker and easier to use.
• Write simply and clearly in plain English to help people get the most value from your website and to help them avoid making mistakes that waste time and energy to resolve.
• Fix any usability issues you identified during your benchmarking to ensure that your website is as easy to use and useful as possible.
• Ensure your user journey is as accessible as possible so the widest possible audience can benefit from using it, offsetting the environmental cost of providing the website.

Step 5: Track And Share Your Progress

As you decarbonise your user journeys, use the benchmarking tools from step 2 to track your progress against the targets you set in step 3 and share your progress as part of your wider sustainability reporting initiatives.

All being well at this point, you will have the numbers to demonstrate how the performance of your user journey has improved and also how you have managed to reduce its carbon footprint.

Share these results with the business as soon as you have them to help you secure the resources to continue the work and initiate similar work on other high-value user journeys.

You should also start to communicate your progress with your users.

It’s important that they are made aware of the carbon footprint of their digital activity and empowered to make informed choices about the environmental impact of the websites that they use.

Ideally, every website should communicate the emissions generated from viewing their pages to help people make these informed choices and also to encourage website providers to minimise their emissions if they are being displayed publicly.

Often, people will have no choice but to use a specific website to complete a specific task, so it is the responsibility of the website provider to ensure the environmental impact of using their website is as small as possible.

You can also help to raise awareness of the environmental impact of websites and what you are doing to minimise your own impact by publishing a digital sustainability statement, such as Unilever’s, as shown below.

A good digital sustainability statement should acknowledge the environmental impact of your website, what you have done to reduce it, and what you plan to do next to minimise it further.

As an industry, we should normalise publishing digital sustainability statements in the same way that accessibility statements have become a standard addition to website footers.

“

Useful Decarbonising Principles

Keep these principles in mind to help you decarbonise your user journeys:

• More doing and less talking.
  Start decarbonising your user journeys as soon as possible to accelerate your learning and positive change.
• Start small.
  Starting small by decarbonising an individual journey makes it easier to get started and generates results to demonstrate value faster.
• Aim to do more with less.
  Minimise what you offer to ensure you are providing the maximum amount of value for the energy you are consuming.
• Make your website as useful and as easy to use as possible.
  Useful websites can justify the energy they consume to provide them, ensuring they are net positive in terms of doing more good than harm.
• Focus on progress over perfection.
  Websites are never finished or perfect but they can always be improved, every small improvement you make will make a difference.

Start Decarbonising Your User Journeys Today

Decarbonising user journeys shouldn’t be done as a one-off, reserved for the next time that you decide to redesign or replatform your website; it should happen on a continual basis as part of your broader digital sustainability strategy.

We know that websites are never finished and that the best websites continually improve as both user and business needs change. I’d like to encourage people to adopt the same mindset when it comes to minimising the environmental impact of their websites.

Decarbonising will happen most effectively when digital professionals challenge themselves on a daily basis to ‘minimise’ the things they are working on.

“

This avoids building ‘carbon debt’ that consists of compounding technical and design debt within our websites, which is always harder to retrospectively remove than avoid in the first place.

By taking a pragmatic approach, such as optimising high-value user journeys and aligning with business metrics such as performance, we stand the best possible chance of making digital sustainability a priority.

You’ll have noticed that, other than using website carbon calculator tools, this approach doesn’t require any skills that don’t already exist within typical digital teams today. This is great because it means you’ve already got the skills that you need to do this important work.

I would encourage everyone to raise the issue of the environmental impact of the internet in their next team meeting and to try this decarbonising approach to create better outcomes for people, profit, performance, purpose, and the planet.

Good luck!

(yk)

Traditional personas suck for UX work. They obsess over marketing metrics like age, income, and job titles while missing what actually matters in design: what people are trying to accomplish.

Functional personas, on the other hand, focus on what people are trying to do, not who they are on paper. With a simple AI‑assisted workflow, you can build and maintain personas that actually guide design, content, and conversion decisions.

• Keep users front of mind with task‑driven personas,
• Skip fragile demographics; center on goals, questions, and blockers,
• Use AI to process your messy inputs fast and fill research gaps,
• Validate lightly, ship confidently, and keep them updated.

In this article, I want to breathe new life into a stale UX asset.

For too long, personas have been something that many of us just created, despite the considerable work that goes into them, only to find they have limited usefulness.

I know that many of you may have given up on them entirely, but I am hoping in this post to encourage you that it is possible to create truly useful personas in a lightweight way.

Why Personas Still Matter

Personas give you a shared lens. When everyone uses the same reference point, you cut debate and make better calls. For UX designers, developers, and digital teams, that shared lens keeps you from designing in silos and helps you prioritize work that genuinely improves the experience.

I use personas as a quick test: Would this change help this user complete their task faster, with fewer doubts? If the answer is no (or a shrug), it’s probably a sign the idea isn’t worth pursuing.

From Demographics To Function

Traditional personas tell you someone’s age, job title, or favorite brand. That makes a nice poster, but it rarely changes design or copy.

Functional personas flip the script. They describe:

• Goals & tasks: What the person is here to achieve.
• Questions & objections: What they need to know before they act.
• Touchpoints: How the person interacts with the organization.
• Service gaps: How the company might be letting this persona down.

When you center on tasks and friction, you get direct lines from user needs to UI decisions, content, and conversion paths.

But remember, this list isn’t set in stone — adapt it to what’s actually useful in your specific situation.

One of the biggest problems with traditional personas was following a rigid template regardless of whether it made sense for your project. We must not fall into that same mistake with functional personas.

“

The Benefits of Functional Personas

For small startups, functional personas reduce wasted effort. For enterprise teams, they keep sprawling projects grounded in what matters most.

However, because of the way we are going to produce our personas, they provide certain benefits in either case:

• Lighten the load: They’re easier to update without large research cycles.
• Stay current: Because they are easy to produce, we can update them more often.
• Tie to outcomes: Tasks, objections, and proof points map straight to funnels, flows, and product decisions.

We can deliver these benefits because we are going to use AI to help us, rather than carrying out a lot of time-consuming new research.

How AI Helps Us Get There

Of course, doing fresh research is always preferable. But in many cases, it is not feasible due to time or budget constraints. I would argue that using AI to help us create personas based on existing assets is preferable to having no focus on user attention at all.

AI tools can chew through the inputs you already have (surveys, analytics, chat logs, reviews) and surface patterns you can act on. They also help you scan public conversations around your product category to fill gaps.

I therefore recommend using AI to:

• Synthesize inputs: Turn scattered notes into clean themes.
• Spot segments by need: Group people by jobs‑to‑be‑done, not demographics.
• Draft quickly: Produce first‑pass personas and sample journeys in minutes.
• Iterate with stakeholders: Update on the fly as you get feedback.

AI doesn’t remove the need for traditional research. Rather, it is a way of extracting more value from the scattered insights into users that already exist within an organization or online.

“

The Workflow

Here’s how to move from scattered inputs to usable personas. Each step builds on the last, so treat it as a cycle you can repeat as projects evolve.

1. Set Up A Dedicated Workspace

Create a dedicated space within your AI tool for this work. Most AI platforms offer project management features that let you organize files and conversations:

• In ChatGPT and Claude, use “Projects” to store context and instructions.
• In Perplexity, Gemini and CoPilot similar functionality is referred to as “Spaces.”

This project space becomes your central repository where all uploaded documents, research data, and generated personas live together. The AI will maintain context between sessions, so you won’t have to re-upload materials each time you iterate. This structured approach makes your workflow more efficient and helps the AI deliver more consistent results.

2. Write Clear Instructions

Next, you can brief your AI project so that it understands what it wants from you. For example:

“Act as a user researcher. Create realistic, functional personas using the project files and public research. Segment by needs, tasks, questions, pain points, and goals. Show your reasoning.”

Asking for a rationale gives you a paper trail you can defend to stakeholders.

3. Upload What You’ve Got (Even If It’s Messy)

This is where things get really powerful.

Upload everything (and I mean everything) you can put your hands on relating to the user. Old surveys, past personas, analytics screenshots, FAQs, support tickets, review snippets; dump them all in. The more varied the sources, the stronger the triangulation.

4. Run Focused External Research

Once you have done that, you can supplement that data by getting AI to carry out “deep research” about your brand. Have AI scan recent (I often focus on the last year) public conversations for your brand, product space, or competitors. Look for:

• Who’s talking and what they’re trying to do;
• Common questions and blockers;
• Phrases people use (great for copywriting).

Save the report you get back into your project.

5. Propose Segments By Need

Once you have done that, ask AI to suggest segments based on tasks and friction points (not demographics). Push back until each segment is distinct, observable, and actionable. If two would behave the same way in your flow, merge them.

This takes a little bit of trial and error and is where your experience really comes into play.

6. Generate Draft Personas

Now you have your segments, the next step is to draft your personas. Use a simple template so the document is read and used. If your personas become too complicated, people will not read them. Each persona should:

• State goals and tasks,
• List objections and blockers,
• Highlight pain points,
• Show touchpoints,
• Identify service gaps.

Below is a sample template you can work with:

# Persona Title: e.g. Savvy Shopper
- Person's Name: e.g. John Smith.
- Age: e.g. 24
- Job: e.g. Social Media Manager

"A quote that sums up the persona's general attitude"

## Primary Goal
What they’re here to achieve (1–2 lines).

## Key Tasks
• Task 1
• Task 2
• Task 3

## Questions & Objections
• What do they need to know before they act?
• What might make them hesitate?

## Pain Points
• Where do they get stuck?
• What feels risky, slow, or confusing?

## Touchpoints
• What channels are they most commonly interacting with?

## Service Gaps
• How is the organization currently failing this persona?

Remember, you should customize this to reflect what will prove useful within your organization.

7. Validate

It is important to validate that what the AI has produced is realistic. Obviously, no persona is a true representation as it is a snapshot in time of a Hypothetical user. However, we do want it to be as accurate as possible.

Share your drafts with colleagues who interact regularly with real users — people in support cells or research teams. Where possible, test with a handful of users. Then cut anything that you can’t defend or correct any errors that are identified.

Troubleshooting & Guardrails

As you work through the above process, you will encounter problems. Here are common pitfalls and how to avoid them:

• Too many personas?
  Merge until each one changes a design or copy decision. Three strong personas beat seven weak ones.
• Stakeholder wants demographics?
  Only include details that affect behavior. Otherwise, leave them out. Suggest separate personas for other functions (such as marketing).
• AI hallucinations?
  Always ask for a rationale or sources. Cross‑check with your own data and customer‑facing teams.
• Not enough data?
  Mark assumptions clearly, then validate with quick interviews, surveys, or usability tests.

Making Personas Useful In Practice

The most important thing to remember is to actually use your personas once they’ve been created. They can easily become forgotten PDFs rather than active tools. Instead, personas should shape your work and be referenced regularly. Here are some ways you can put personas to work:

• Navigation & IA: Structure menus by top tasks.
• Content & Proof: Map objections to FAQs, case studies, and microcopy.
• Flows & UI: Streamline steps to match how people think.
• Conversion: Match CTAs to personas’ readiness, goals, and pain points.
• Measurement: Track KPIs that map to personas, not vanity metrics.

With this approach, personas evolve from static deliverables into dynamic reference points your whole team can rely on.

Keep Them Alive

Treat personas as a living toolkit. Schedule a refresh every quarter or after major product changes. Rerun the research pass, regenerate summaries, and archive outdated assumptions. The goal isn’t perfection; it’s keeping them relevant enough to guide decisions.

Bottom Line

Functional personas are faster to build, easier to maintain, and better aligned with real user behavior. By combining AI’s speed with human judgment, you can create personas that don’t just sit in a slide deck; they actively shape better products, clearer interfaces, and smoother experiences.

(yk)

In “A Week In The Life Of An AI-Augmented Designer”, we followed Kate’s weeklong journey of her first AI-augmented design sprint. She had three realizations through the process:

1. AI isn’t a co-pilot (yet); it’s more like a smart, eager intern.
   One with access to a lot of information, good recall, fast execution, but no context. That mindset defined how she approached every interaction with AI: not as magic, but as management.
2. Don’t trust; guide, coach, and always verify.
   Like any intern, AI needs coaching and supervision, and that’s where her designerly skills kicked in. Kate relied on curiosity to explore, observation to spot bias, empathy to humanize the output, and critical thinking to challenge what didn’t feel right. Her learning mindset helped her keep up with advances, and experimentation helped her learn by doing.
3. Prompting is part creative brief, and part conversation design, just with an AI instead of a person.
   When you prompt an AI, you’re not just giving instructions, but designing how it responds, behaves, and outputs information. If AI is like an intern, then the prompt is your creative brief that frames the task, sets the tone, and clarifies what good looks like. It’s also your conversation script that guides how it responds, how the interaction flows, and how ambiguity is handled.

As designers, we’re used to designing interactions for people. Prompting is us designing our own interactions with machines — it uses the same mindset with a new medium. It shapes an AI’s behavior the same way you’d guide a user with structure, clarity, and intent.

If you’ve bookmarked, downloaded, or saved prompts from others, you’re not alone. We’ve all done that during our AI journeys. But while someone else’s prompts are a good starting point, you will get better and more relevant results if you can write your own prompts tailored to your goals, context, and style. Using someone else’s prompt is like using a Figma template. It gets the job done, but mastery comes from understanding and applying the fundamentals of design, including layout, flow, and reasoning. Prompts have a structure too. And when you learn it, you stop guessing and start designing.

Note: All prompts in this article were tested using ChatGPT — not because it’s the only game in town, but because it’s friendly, flexible, and lets you talk like a person, yes, even after the recent GPT-5 “update”. That said, any LLM with a decent attention span will work. Results for the same prompt may vary based on the AI model you use, the AI’s training, mood, and how confidently it can hallucinate.

Privacy PSA: As always, don’t share anything you wouldn’t want leaked, logged, or accidentally included in the next AI-generated meme. Keep it safe, legal, and user-respecting.

With that out of the way, let’s dive into the mindset, anatomy, and methods of effective prompting as another tool in your design toolkit.

Mindset: Prompt Like A Designer

As designers, we storyboard journeys, wireframe interfaces to guide users, and write UX copy with intention. However, when prompting AI, we treat it differently: “Summarize these insights”, “Make this better”, “Write copy for this screen”, and then wonder why the output feels generic, off-brand, or just meh. It’s like expecting a creative team to deliver great work from a one-line Slack message. We wouldn’t brief a freelancer, much less an intern, with “Design a landing page,” so why brief AI that way?

Prompting Is A Creative Brief For A Machine

Think of a good prompt as a creative brief, just for a non-human collaborator. It needs similar elements, including a clear role, defined goal, relevant context, tone guidance, and output expectations. Just as a well-written creative brief unlocks alignment and quality from your team, a well-structured prompt helps the AI meet your expectations, even though it doesn’t have real instincts or opinions.

Prompting Is Also Conversation Design

A good prompt goes beyond defining the task and sets the tone for the exchange by designing a conversation: guiding how the AI interprets, sequences, and responds. You shape the flow of tasks, how ambiguity is handled, and how refinement happens — that’s conversation design.

Anatomy: Structure It Like A Designer

So how do you write a designer-quality prompt? That’s where the W.I.R.E.+F.R.A.M.E. prompt design framework comes in — a UX-inspired framework for writing intentional, structured, and reusable prompts. Each letter represents a key design direction, grounded in the way UX designers already think: Just as a wireframe doesn’t dictate final visuals, this WIRE+FRAME framework doesn’t constrain creativity, but guides the AI with structured information it needs.

“Why not just use a series of back-and-forth chats with AI?”

You can, and many people do. But without structure, AI fills in the gaps on its own, often with vague or generic results. A good prompt upfront saves time, reduces trial and error, and improves consistency. And whether you’re working on your own or across a team, a framework means you’re not reinventing a prompt every time but reusing what works to get better results faster.

Just as we build wireframes before adding layers of fidelity, the WIRE+FRAME framework has two parts:

• WIRE is the must-have skeleton. It gives the prompt its shape.
• FRAME is the set of enhancements that bring polish, logic, tone, and reusability — like building a high-fidelity interface from the wireframe.

Let’s improve Kate’s original research synthesis prompt (“Read this customer feedback and tell me how we can improve financial literacy for Gen Z in our app”). To better reflect how people actually prompt in practice, let’s tweak it to a more broadly applicable version: “Read this customer feedback and tell me how we can improve our app for Gen Z users.” This one-liner mirrors the kinds of prompts we often throw at AI tools: short, simple, and often lacking structure.

Now, we’ll take that prompt and rebuild it using the first four elements of the W.I.R.E. framework — the core building blocks that provide AI with the main information it needs to deliver useful results.

W: Who & What

Define who the AI should be, and what it’s being asked to deliver.

A creative brief starts with assigning the right hat. Are you briefing a copywriter? A strategist? A product designer? The same logic applies here. Give the AI a clear identity and task. Treat AI like a trusted freelancer or intern. Instead of saying “help me”, tell it who it should act as and what’s expected.

Example: “You are a senior UX researcher and customer insights analyst. You specialize in synthesizing qualitative data from diverse sources to identify patterns, surface user pain points, and map them across customer journey stages. Your outputs directly inform product, UX, and service priorities.”

I: Input Context

Provide background that frames the task.

Creative partners don’t work in a vacuum. They need context: the audience, goals, product, competitive landscape, and what’s been tried already. This is the “What you need to know before you start” section of the brief. Think: key insights, friction points, business objectives. The same goes for your prompt.

Example: “You are analyzing customer feedback for Fintech Brand’s app, targeting Gen Z users. Feedback will be uploaded from sources such as app store reviews, survey feedback, and usability test transcripts.”

R: Rules & Constraints

Clarify any limitations, boundaries, and exclusions.

Good creative briefs always include boundaries — what to avoid, what’s off-brand, or what’s non-negotiable. Things like brand voice guidelines, legal requirements, or time and word count limits. Constraints don’t limit creativity — they focus it. AI needs the same constraints to avoid going off the rails.

Example: “Only analyze the uploaded customer feedback data. Do not fabricate pain points, representative quotes, journey stages, or patterns. Do not supplement with prior knowledge or hypothetical examples. Use clear, neutral, stakeholder-facing language.”

E: Expected Output

Spell out what the deliverable should look like.

This is the deliverable spec: What does the finished product look like? What tone, format, or channel is it for? Even if the task is clear, the format often isn’t. Do you want bullet points or a story? A table or a headline? If you don’t say, the AI will guess, and probably guess wrong. Even better, include an example of the output you want, an effective way to help AI know what you’re expecting. If you’re using GPT-5, you can also mix examples across formats (text, images, tables) together.

Example: “Return a structured list of themes. For each theme, include:

• Theme Title
• Summary of the Issue
• Problem Statement
• Opportunity
• Representative Quotes (from data only)
• Journey Stage(s)
• Frequency (count from data)
• Severity Score (1–5) where 1 = Minor inconvenience or annoyance; 3 = Frustrating but workaround exists; 5 = Blocking issue
• Estimated Effort (Low / Medium / High), where Low = Copy or content tweak; Medium = Logic/UX/UI change; High = Significant changes.”

WIRE gives you everything you need to stop guessing and start designing your prompts with purpose. When you start with WIRE, your prompting is like a briefing, treating AI like a collaborator.

Once you’ve mastered this core structure, you can layer in additional fidelity, like tone, step-by-step flow, or iterative feedback, using the FRAME elements. These five elements provide additional guidance and clarity to your prompt by layering clear deliverables, thoughtful tone, reusable structure, and space for creative iteration.

F: Flow of Tasks

Break complex prompts into clear, ordered steps.

This is your project plan or creative workflow that lays out the stages, dependencies, or sequence of execution. When the task has multiple parts, don’t just throw it all into one sentence. You are doing the thinking and guiding AI. Structure it like steps in a user journey or modules in a storyboard. In this example, it fits as the blueprint for the AI to use to generate the table described in “E: Expected Output”

Example: “Recommended flow of tasks:
Step 1: Parse the uploaded data and extract discrete pain points.
Step 2: Group them into themes based on pattern similarity.
Step 3: Score each theme by frequency (from data), severity (based on content), and estimated effort.
Step 4: Map each theme to the appropriate customer journey stage(s).
Step 5: For each theme, write a clear problem statement and opportunity based only on what’s in the data.”

R: Reference Voice or Style

Name the desired tone, mood, or reference brand.

This is the brand voice section or style mood board — reference points that shape the creative feel. Sometimes you want buttoned-up. Other times, you want conversational. Don’t assume the AI knows your tone, so spell it out.

Example: “Use the tone of a UX insights deck or product research report. Be concise, pattern-driven, and objective. Make summaries easy to scan by product managers and design leads.”

A: Ask for Clarification

Invite the AI to ask questions before generating, if anything is unclear.

This is your “Any questions before we begin?” moment — a key step in collaborative creative work. You wouldn’t want a freelancer to guess what you meant if the brief was fuzzy, so why expect AI to do better? Ask AI to reflect or clarify before jumping into output mode.

Example: “If the uploaded data is missing or unclear, ask for it before continuing. Also, ask for clarification if the feedback format is unstructured or inconsistent, or if the scoring criteria need refinement.”

M: Memory (Within The Conversation)

Reference earlier parts of the conversation and reuse what’s working.

This is similar to keeping visual tone or campaign language consistent across deliverables in a creative brief. Prompts are rarely one-shot tasks, so this reminds AI of the tone, audience, or structure already in play. GPT-5 got better with memory, but this still remains a useful element, especially if you switch topics or jump around.

Example: “Unless I say otherwise, keep using this process: analyze the data, group into themes, rank by importance, then suggest an action for each.”

E: Evaluate & Iterate

Invite the AI to critique, improve, or generate variations.

This is your revision loop — your way of prompting for creative direction, exploration, and refinement. Just like creatives expect feedback, your AI partner can handle review cycles if you ask for them. Build iteration into the brief to get closer to what you actually need. Sometimes, you may see ChatGPT test two versions of a response on its own by asking for your preference.

Example: “After listing all themes, identify the one with the highest combined priority score (based on frequency, severity, and effort).

For that top-priority theme:

• Critically evaluate its framing: Is the title clear? Are the quotes strong and representative? Is the journey mapping appropriate?
• Suggest one improvement (e.g., improved title, more actionable implication, clearer quote, tighter summary).
• Rewrite the theme entry with that improvement applied.
• Briefly explain why the revision is stronger and more useful for product or design teams.”

Here’s a quick recap of the WIRE+FRAME framework:

And here’s the full WIRE+FRAME prompt:

(W) You are a senior UX researcher and customer insights analyst. You specialize in synthesizing qualitative data from diverse sources to identify patterns, surface user pain points, and map them across customer journey stages. Your outputs directly inform product, UX, and service priorities.

(I) You are analyzing customer feedback for Fintech Brand’s app, targeting Gen Z users. Feedback will be uploaded from sources such as app store reviews, survey feedback, and usability test transcripts.

(R) Only analyze the uploaded customer feedback data. Do not fabricate pain points, representative quotes, journey stages, or patterns. Do not supplement with prior knowledge or hypothetical examples. Use clear, neutral, stakeholder-facing language.

(E) Return a structured list of themes. For each theme, include:

• Theme Title
• Summary of the Issue
• Problem Statement
• Opportunity
• Representative Quotes (from data only)
• Journey Stage(s)
• Frequency (count from data)
• Severity Score (1–5) where 1 = Minor inconvenience or annoyance; 3 = Frustrating but workaround exists; 5 = Blocking issue
• Estimated Effort (Low / Medium / High), where Low = Copy or content tweak; Medium = Logic/UX/UI change; High = Significant changes

(F) Recommended flow of tasks:
Step 1: Parse the uploaded data and extract discrete pain points.
Step 2: Group them into themes based on pattern similarity.
Step 3: Score each theme by frequency (from data), severity (based on content), and estimated effort.
Step 4: Map each theme to the appropriate customer journey stage(s).
Step 5: For each theme, write a clear problem statement and opportunity based only on what’s in the data.

(R) Use the tone of a UX insights deck or product research report. Be concise, pattern-driven, and objective. Make summaries easy to scan by product managers and design leads.

(A) If the uploaded data is missing or unclear, ask for it before continuing. Also, ask for clarification if the feedback format is unstructured or inconsistent, or if the scoring criteria need refinement.

(M) Unless I say otherwise, keep using this process: analyze the data, group into themes, rank by importance, then suggest an action for each.

(E) After listing all themes, identify the one with the highest combined priority score (based on frequency, severity, and effort).
For that top-priority theme:

• Critically evaluate its framing: Is the title clear? Are the quotes strong and representative? Is the journey mapping appropriate?
• Suggest one improvement (e.g., improved title, more actionable implication, clearer quote, tighter summary).
• Rewrite the theme entry with that improvement applied.
• Briefly explain why the revision is stronger and more useful for product or design teams.

You could use “##” to label the sections (e.g., “##FLOW”) more for your readability than for AI. At over 400 words, this Insights Synthesis prompt example is a detailed, structured prompt, but it isn’t customized for you and your work. The intent wasn’t to give you a specific prompt (the proverbial fish), but to show how you can use a prompt framework like WIRE+FRAME to create a customized, relevant prompt that will help AI augment your work (teaching you to fish).

Keep in mind that prompt length isn’t a common concern, but rather a lack of quality and structure is. As of the time of writing, AI models can easily process prompts that are thousands of words long.

Not every prompt needs all the FRAME components; WIRE is often enough to get the job done. But when the work is strategic or highly contextual, pick components from FRAME — the extra details can make a difference. Together, WIRE+FRAME give you a detailed framework for creating a well-structured prompt, with the crucial components first, followed by optional components:

• WIRE builds a clear, focused prompt with role, input, rules, and expected output.
• FRAME adds refinement like tone, reusability, and iteration.

Here are some scenarios and recommendations for using WIRE or WIRE+FRAME:

Prompting isn’t about getting it right the first time. It’s about designing the interaction and redesigning when needed. With WIRE+FRAME, you’re going beyond basic prompting and designing the interaction between you and AI.

From Gut Feel To Framework: A Prompt Makeover

Let’s compare the results of Kate’s first AI-augmented design sprint prompt (to synthesize customer feedback into design insights) with one based on the WIRE+FRAME prompt framework, with the same data and focusing on the top results:

Original prompt: Read this customer feedback and tell me how we can improve our app for Gen Z users.

Initial ChatGPT Results:

• Improve app reliability to reduce crashes and freezing.
• Provide better guidance or tutorials for financial tools like budgeting or goal setting.
• Enhance the transparency of Zelle transfers by showing confirmation messages.
• Speed up app loading and reduce lag on key actions.

With this version, you’d likely need to go back and forth with follow-up questions, rewrite the output for clarity, and add structure before sharing with your team.

WIRE+FRAME prompt above (with defined role, scope, rules, expected format, tone, flow, and evaluation loop).

Initial ChatGPT Results:

You can clearly see the very different results from the two prompts, both using the exact same data. While the first prompt returns a quick list of ideas, the detailed WIRE+FRAME version doesn’t just summarize feedback but structures it. Themes are clearly labeled, supported by user quotes, mapped to customer journey stages, and prioritized by frequency, severity, and effort.

The structured prompt results can be used as-is or shared without needing to reformat, rewrite, or explain them (see disclaimer below). The first prompt output needs massaging: it’s not detailed, lacks evidence, and would require several rounds of clarification to be actionable. The first prompt may work when the stakes are low and you are exploring. But when your prompt is feeding design, product, or strategy, structure comes to the rescue.

Disclaimer: Know Your Data

A well-structured prompt can make AI output more useful, but it shouldn’t be the final word, or your single source of truth. AI models are powerful pattern predictors, not fact-checkers. If your data is unclear or poorly referenced, even the best prompt may return confident nonsense. Don’t blindly trust what you see. Treat AI like a bright intern: fast, eager, and occasionally delusional. You should always be familiar with your data and validate what AI spits out. For example, in the WIRE+FRAME results above, AI rated the effort as low for financial tool onboarding. That could easily be a medium or high. Good prompting should be backed by good judgment.

Try This Now

Start by using the WIRE+FRAME framework to create a prompt that will help AI augment your work. You could also rewrite the last prompt you were not satisfied with, using the WIRE+FRAME, and compare the output.

Feel free to use this simple tool to guide you through the framework.

Methods: From Lone Prompts to a Prompt System

Just as design systems have reusable components, your prompts can too. You can use the WIRE+FRAME framework to write detailed prompts, but you can also use the structure to create reusable components that are pre-tested, plug-and-play pieces you can assemble to build high-quality prompts faster. Each part of WIRE+FRAME can be transformed into a prompt component: small, reusable modules that reflect your team’s standards, voice, and strategy.

For instance, if you find yourself repeatedly using the same content for different parts of the WIRE+FRAME framework, you could save them as reusable components for you and your team. In the example below, we have two different reusable components for “W: Who & What” — an insights analyst and an information architect.

W: Who & What

1. You are a senior UX researcher and customer insights analyst. You specialize in synthesizing qualitative data from diverse sources to identify patterns, surface user pain points, and map them across customer journey stages. Your outputs directly inform product, UX, and service priorities.
2. You are an experienced information architect specializing in organizing enterprise content on intranets. Your task is to reorganize the content and features into categories that reflect user goals, reduce cognitive load, and increase findability.

Create and save prompt components and variations for each part of the WIRE+FRAME framework, allowing your team to quickly assemble new prompts by combining components when available, rather than starting from scratch each time.

Behind The Prompts: Questions About Prompting

Q: If I use a prompt framework like WIRE+FRAME every time, will the results be predictable?

A: Yes and no. Yes, your outputs will be guided by a consistent set of instructions (e.g., Rules, Examples, Reference Voice / Style) that will guide the AI to give you a predictable format and style of results. And no, while the framework provides structure, it doesn’t flatten the generative nature of AI, but focuses it on what’s important to you. In the next article, we will look at how you can use this to your advantage to quickly reuse your best repeatable prompts as we build your AI assistant.

Q: Could changes to AI models break the WIRE+FRAME framework?

A: AI models are evolving more rapidly than any other technology we’ve seen before — in fact, ChatGPT was recently updated to GPT-5 to mixed reviews. The update didn’t change the core principles of prompting or the WIRE+FRAME prompt framework. With future releases, some elements of how we write prompts today may change, but the need to communicate clearly with AI won’t. Think of how you delegate work to an intern vs. someone with a few years’ experience: you still need detailed instructions the first time either is doing a task, but the level of detail may change. WIRE+FRAME isn’t built only for today’s models; the components help you clarify your intent, share relevant context, define constraints, and guide tone and format — all timeless elements, no matter how smart the model becomes. The skill of shaping clear, structured interactions with non-human AI systems will remain valuable.

Q: Can prompts be more than text? What about images or sketches?

A: Absolutely. With tools like GPT-5 and other multimodal models, you can upload screenshots, pictures, whiteboard sketches, or wireframes. These visuals become part of your Input Context or help define the Expected Output. The same WIRE+FRAME principles still apply: you’re setting context, tone, and format, just using images and text together. Whether your input is a paragraph or an image and text, you’re still designing the interaction.

Have a prompt-related question of your own? Share it in the comments, and I’ll either respond there or explore it further in the next article in this series.

From Designerly Prompting To Custom Assistants

Good prompts and results don’t come from using others’ prompts, but from writing prompts that are customized for you and your context. The WIRE+FRAME framework helps with that and makes prompting a tool you can use to guide AI models like a creative partner instead of hoping for magic from a one-line request.

Prompting uses the designerly skills you already use every day to collaborate with AI:

• Curiosity to explore what the AI can do and frame better prompts.
• Observation to detect bias or blind spots.
• Empathy to make machine outputs human.
• Critical thinking to verify and refine.
• Experiment & Iteration to learn by doing and improve the interaction over time.
• Growth Mindset to keep up with new technology like AI and prompting.

Once you create and refine prompt components and prompts that work for you, make them reusable by documenting them. But wait, there’s more — what if your best prompts, or the elements of your prompts, could live inside your own AI assistant, available on demand, fluent in your voice, and trained on your context? That’s where we’re headed next.

In the next article, “Design Your Own Design Assistant”, we’ll take what you’ve learned so far and turn it into a Custom AI assistant (aka Custom GPT), a design-savvy, context-aware assistant that works like you do. We’ll walk through that exact build, from defining the assistant’s job description to uploading knowledge, testing, and sharing it with others.

Resources

• GPT-5 Prompting Guide
• GPT-4.1 Prompting Guide
• Anthropic Prompt Engineering
• Prompt Engineering by Google
• Perplexity
• Webapp to guide you through the WIRE+FRAME framework

(yk)

Artificial Intelligence isn’t new, but in November 2022, something changed. The launch of ChatGPT brought AI out of the background and into everyday life. Suddenly, interacting with a machine didn’t feel technical — it felt conversational.

Just this March, ChatGPT overtook Instagram and TikTok as the most downloaded app in the world. That level of adoption shows that millions of everyday users, not just developers or early adopters, are comfortable using AI in casual, conversational ways. People are using AI not just to get answers, but to think, create, plan, and even to help with mental health and loneliness.

In the past two and a half years, people have moved through the Kübler-Ross Change Curve — only instead of grief, it’s AI-induced uncertainty. UX designers, like Kate (who you’ll meet shortly), have experienced something like this:

• Denial: “AI can’t design like a human; it won’t affect my workflow.”
• Anger: “AI will ruin creativity. It’s a threat to our craft.”
• Bargaining: “Okay, maybe just for the boring tasks.”
• Depression: “I can’t keep up. What’s the future of my skills?”
• Acceptance: “Alright, AI can free me up for more strategic, human work.”

As designers move into experimentation, they’re not asking, Can I use AI? but How might I use it well?.

Using AI isn’t about chasing the latest shiny object but about learning how to stay human in a world of machines, and use AI not as a shortcut, but as a creative collaborator.

“

It isn’t about finding, bookmarking, downloading, or hoarding prompts, but experimenting and writing your own prompts.

To bring this to life, we’ll follow Kate, a mid-level designer at a FinTech company, navigating her first AI-augmented design sprint. You’ll see her ups and downs as she experiments with AI, tries to balance human-centered skills with AI tools, when she relies on intuition over automation, and how she reflects critically on the role of AI at each stage of the sprint.

The next two planned articles in this series will explore how to design prompts (Part 2) and guide you through building your own AI assistant (aka CustomGPT; Part 3). Along the way, we’ll spotlight the designerly skills AI can’t replicate like curiosity, empathy, critical thinking, and experimentation that will set you apart in a world where automation is easy, but people and human-centered design matter even more.

Note: This article was written by a human (with feelings, snacks, and deadlines). The prompts are real, the AI replies are straight from the source, and no language models were overworked — just politely bossed around. All em dashes are the handiwork of MS Word’s autocorrect — not AI. Kate is fictional, but her week is stitched together from real tools, real prompts, real design activities, and real challenges designers everywhere are navigating right now. She will primarily be using ChatGPT, reflecting the popularity of this jack-of-all-trades AI as the place many start their AI journeys before branching out. If you stick around to the end, you’ll find other AI tools that may be better suited for different design sprint activities. Due to the pace of AI advances, your outputs may vary (YOMV), possibly by the time you finish reading this sentence.

Cautionary Note: AI is helpful, but not always private or secure. Never share sensitive, confidential, or personal information with AI tools — even the helpful-sounding ones. When in doubt, treat it like a coworker who remembers everything and may not be particularly good at keeping secrets.

Prologue: Meet Kate (As She Preps For The Upcoming Week)

Kate stared at the digital mountain of feedback on her screen: transcripts, app reviews, survey snippets, all waiting to be synthesized. Deadlines loomed. Her calendar was a nightmare. Meanwhile, LinkedIn was ablaze with AI hot takes and success stories. Everyone seemed to have found their “AI groove” — except her. She wasn’t anti-AI. She just hadn’t figured out how it actually fit into her work. She had tried some of the prompts she saw online, played with some AI plugins and extensions, but it felt like an add-on, not a core part of her design workflow.

Her team was focusing on improving financial confidence for Gen Z users of their FinTech app, and Kate planned to use one of her favorite frameworks: the Design Sprint, a five-day, high-focus process that condenses months of product thinking into a single week. Each day tackles a distinct phase: Understand, Sketch, Decide, Prototype, and Test. All designed to move fast, make ideas tangible, and learn from real users before making big bets.

This time, she planned to experiment with a very lightweight version of the design sprint, almost “solo-ish” since her PM and engineer were available for check-ins and decisions, but not present every day. That gave her both space and a constraint, and made it the perfect opportunity to explore how AI could augment each phase of the sprint.

She decided to lean on her designerly behavior of experimentation and learning and integrate AI intentionally into her sprint prep, using it as both a creative partner and a thinking aid. Not with a rigid plan, but with a working hypothesis that AI would at the very least speed her up, if nothing else.

She wouldn’t just be designing and testing a prototype, but prototyping and testing what it means to design with AI, while still staying in the driver’s seat.

Follow Kate along her journey through her first AI-powered design sprint: from curiosity to friction and from skepticism to insight.

Monday: Understanding the Problem (aka: Kate Vs. Digital Pile Of Notes)

The first day of a design sprint is spent understanding the user, their problems, business priorities, and technical constraints, and narrowing down the problem to solve that week.

This morning, Kate had transcripts from recent user interviews and customer feedback from the past year from app stores, surveys, and their customer support center. Typically, she would set aside a few days to process everything, coming out with glazed eyes and a few new insights. This time, she decided to use ChatGPT to summarize that data: “Read this customer feedback and tell me how we can improve financial literacy for Gen Z in our app.”

ChatGPT’s outputs were underwhelming to say the least. Disappointed, she was about to give up when she remembered an infographic about good prompting that she had emailed herself. She updated her prompt based on those recommendations:

• Defined a role for the AI (“product strategist”),
• Provided context (user group and design sprint objectives), and
• Clearly outlined what she was looking for (financial literacy related patterns in pain points, blockers, confusion, lack of confidence; synthesis to identify top opportunity areas).

By the time she Aero-pressed her next cup of coffee, ChatGPT had completed its analysis, highlighting blockers like jargon, lack of control, fear of making the wrong choice, and need for blockchain wallets. Wait, what? That last one felt off.

Kate searched her sources and confirmed her hunch: AI hallucination! Despite the best of prompts, AI sometimes makes things up based on trendy concepts from its training data rather than actual data. Kate updated her prompt with constraints to make ChatGPT only use data she had uploaded, and to cite examples from that data in its results. 18 seconds later, the updated results did not mention blockchain or other unexpected results.

By lunch, Kate had the makings of a research summary that would have taken much, much longer, and a whole lot of caffeine.

That afternoon, Kate and her product partner plotted the pain points on the Gen Z app journey. The emotional mapping highlighted the most critical moment: the first step of a financial decision, like setting a savings goal or choosing an investment option. That was when fear, confusion, and lack of confidence held people back.

AI synthesis combined with human insight helped them define the problem statement as: “How might we help Gen Z users confidently take their first financial action in our app, in a way that feels simple, safe, and puts them in control?”

Kate’s Reflection

As she wrapped up for the day, Kate jotted down her reflections on her first day as an AI-augmented designer:

There’s nothing like learning by doing. I’ve been reading about AI and tinkering around, but took the plunge today. Turns out AI is much more than a tool, but I wouldn’t call it a co-pilot. Yet. I think it’s like a sharp intern: it has a lot of information, is fast, eager to help, but it lacks context, needs supervision, and can surprise you. You have to give it clear instructions, double-check its work, and guide and supervise it. Oh, and maintain boundaries by not sharing anything I wouldn’t want others to know.

Today was about listening — to users, to patterns, to my own instincts. AI helped me sift through interviews fast, but I had to stay curious to catch what it missed. Some quotes felt too clean, like the edges had been smoothed over. That’s where observation and empathy kicked in. I had to ask myself: what’s underneath this summary?

Critical thinking was the designerly skill I had to exercise most today. It was tempting to take the AI’s synthesis at face value, but I had to push back by re-reading transcripts, questioning assumptions, and making sure I wasn’t outsourcing my judgment. Turns out, the thinking part still belongs to me.

Tuesday: Sketching (aka: Kate And The Sea of OKish Ideas)

Day 2 of a design sprint focuses on solutions, starting by remixing and improving existing ideas, followed by people sketching potential solutions.

Optimistic, yet cautious after her experience yesterday, Kate started thinking about ways she could use AI today, while brewing her first cup of coffee. By cup two, she was wondering if AI could be a creative teammate. Or a creative intern at least. She decided to ask AI for a list of relevant UX patterns across industries. Unlike yesterday’s complex analysis, Kate was asking for inspiration, not insight, which meant she could use a simpler prompt: “Give me 10 unique examples of how top-rated apps reduce decision anxiety for first-time users — from FinTech, health, learning, or ecommerce.”

She received her results in a few seconds, but there were only 6, not the 10 she asked for. She expanded her prompt for examples from a wider range of industries. While reviewing the AI examples, Kate realized that one had accessibility issues. To be fair, the results met Kate’s ask since she had not specified accessibility considerations. She then went pre-AI and brainstormed examples with her product partner, coming up with a few unique local examples.

Later that afternoon, Kate went full human during Crazy 8s by putting a marker to paper and sketching 8 ideas in 8 minutes to rapidly explore different directions. Wondering if AI could live up to its generative nature, she uploaded pictures of her top 3 sketches and prompted AI to act as “a product design strategist experienced in Gen Z behavior, digital UX, and behavioral science”, gave it context about the problem statement, stage in the design sprint, and explicitly asked AI the following:

1. Analyze the 3 sketch concepts and identify core elements or features that resonated with the goal.
2. Generate 5 new concept directions, each of which should:
   • Address the original design sprint challenge.
   • Reflect Gen Z design language, tone, and digital behaviors.
   • Introduce a unique twist, remix, or conceptual inversion of the ideas in the sketches.
3. For each concept, provide:
   • Name (e.g., “Monopoly Mode,” “Smart Start”);
   • 1–2 sentence concept summary;
   • Key differentiator from the original sketches;
   • Design tone and/or behavioral psychology technique used.

The results included ideas that Kate and her product partner hadn’t considered, including a progress bar that started at 20% (to build confidence), and a sports-like “stock bracket” for first-time investors.

Not bad, thought Kate, as she cherry-picked elements, combined and built on these ideas in her next round of sketches. By the end of the day, they had a diverse set of sketched solutions — some original, some AI-augmented, but all exploring how to reduce fear, simplify choices, and build confidence for Gen Z users taking their first financial step. With five concept variations and a few rough storyboards, Kate was ready to start converging on day 3.

Kate’s Reflection

Today was creatively energizing yet a little overwhelming! I leaned hard on AI to act as a creative teammate. It delivered a few unexpected ideas and remixed my Crazy 8s into variations I never would’ve thought of!

It also reinforced the need to stay grounded in the human side of design. AI was fast — too fast, sometimes. It spit out polished-sounding ideas that sounded right, but I had to slow down, observe carefully, and ask: Does this feel right for our users? Would a first-time user feel safe or intimidated here?

Critical thinking helped me separate what mattered from what didn’t. Empathy pulled me back to what Gen Z users actually said, and kept their voices in mind as I sketched. Curiosity and experimentation were my fuel. I kept tweaking prompts, remixing inputs, and seeing how far I could stretch a concept before it broke. Visual communication helped translate fuzzy AI ideas into something I could react to — and more importantly, test.

Wednesday: Deciding (aka Kate Tries to Get AI to Pick a Side)

Design sprint teams spend Day 3 critiquing each of their potential solutions to shortlist those that have the best chance of achieving their long-term goal. The winning scenes from the sketches are then woven into a prototype storyboard.

Design sprint Wednesdays were Kate’s least favorite day. After all the generative energy during Sketching Tuesday, today, she would have to decide on one clear solution to prototype and test. She was unsure if AI would be much help with judging tradeoffs or narrowing down options, and it wouldn’t be able to critique like a team. Or could it?

Kate reviewed each of the five concepts, noting strengths, open questions, and potential risks. Curious about how AI would respond, she uploaded images of three different design concepts and prompted ChatGPT for strengths and weaknesses. AI’s critique was helpful in summarizing the pros and cons of different concepts, including a few points she had not considered — like potential privacy concerns.

She asked a few follow-up questions to confirm the actual reasoning. Wondering if she could simulate a team critique by prompting ChatGPT differently, Kate asked it to use the 6 thinking hats technique. The results came back dense, overwhelming, and unfocused. The AI couldn’t prioritize, and it couldn’t see the gaps Kate instinctively noticed: friction in onboarding, misaligned tone, unclear next steps.

In that moment, the promise of AI felt overhyped. Kate stood up, stretched, and seriously considered ending her experiments with the AI-driven process. But she paused. Maybe the problem wasn’t the tool. Maybe it was how she was using it. She made a note to experiment when she wasn’t on a design sprint clock.

She returned to her sketches, this time laying them out on the wall. No screens, no prompts. Just markers, sticky notes, and Sharpie scribbles. Human judgment took over. Kate worked with her product partner to finalize the solution to test on Friday and spent the next hour storyboarding the experience in Figma.

Kate re-engaged with AI as a reviewer, not a decider. She prompted it for feedback on the storyboard and was surprised to see it spit out detailed design, content, and micro-interaction suggestions for each of the steps of the storyboarded experience. A lot of food for thought, but she’d have to judge what mattered when she created her prototype. But that wasn’t until tomorrow!

Kate’s Reflection

AI exposed a few of my blind spots in the critique, which was good, but it basically pointed out that multiple options “could work”. I had to rely on my critical thinking and instincts to weigh options logically, emotionally, and contextually in order to choose a direction that was the most testable and aligned with the user feedback from Day 1.

I was also surprised by the suggestions it came up with while reviewing my final storyboard, but I will need a fresh pair of eyes and all the human judgement I can muster tomorrow.

Empathy helped me walk through the flow like I was a new user. Visual communication helped pull it all together by turning abstract steps into a real storyboard for the team to see instead of imagining.

TO DO: Experiment prompting around the 6 Thinking Hats for different perspectives.

Thursday: Prototype (aka Kate And Faking It)

On Day 4, the team usually turns the storyboard from the previous day into a prototype that can be tested with users on Day 5. The prototype doesn’t need to be fully functional; a simulated experience is sufficient to gather user feedback.

Kate’s prototype day often consisted of marathon Figma Design sessions and late-night pizza dinners. She was hoping AI would change that today. She fed yesterday’s storyboard to ChatGPT and asked it for screens. It took a while to generate, but she was excited to see a screen flow gradually appear on her screen, except that it had 3 ¾ screens, instead of the 6 frames from her storyboard, as you can see in the image below.

She tried different prompts but couldn’t get a complete flow. She was excited when it offered to create a Figma file. She waited. And waited. And reminded it that she was right there waiting. It finally gave her a link — a link that didn’t work. That made her question why she needed prototype screens in Figma at all. Could one of those no-code tools do the job? Kate remembered that she had just received beta access to Figma Make (Figma’s AI prompt-to-prototype tool) and decided to put it to the test.

She shared her storyboard, added context about the prototype for a design sprint, the problem, and a brief description of the audience. In a few seconds, she saw code being generated and a running description on the left. It seemed like it would take some time, so she headed to the cafeteria for lunch.

When Kate returned, she was surprised to see Figma Make had created an app along with a detailed explanation:

“I’ve created a complete finance app prototype with five key screens that guide users through setting financial goals in a friendly, low-pressure way. The app features smooth animations with Framer Motion, a progress indicator to show users where they are in the journey, and emotionally supportive visuals and copy throughout the experience. Each screen serves a specific purpose: welcoming users, helping them define relatable goals, building their plan with visual feedback, presenting a simple summary, and offering next steps without pressure.

You might want to:

• Add personalization options like user profiles or avatars.
• Implement a dark mode toggle for accessibility.
• Create additional goal templates and saving strategies.”

Since Figma Make had built the prototype, not her, Kate felt like she was User 0, and clicked her way through the interactive prototype, making a mental note of what worked, what didn’t. She was impressed to see that she could not only select a common goal (buy a home, save for education, and so on) but she could also create her own goal (hello De’Longhi Coffee Maker) and it carried over for the rest of the experience. This was something she had never been able to do in Figma Design!

Despite some obvious misses like a missing header and navigation, and some buttons not working, she was impressed! Kate tried the option to ‘Publish’ and it gave her a link that she immediately shared with her product and engineering partners. A few minutes later, they joined her in the conference room, exploring it together. The engineer scanned the code, didn’t seem impressed, but said it would work as a disposable prototype.

Kate prompted Figma Make to add an orange header and app navigation, and this time the trio kept their eyes peeled as they saw the progress in code and in English. The results were pretty good. They spent the next hour making changes to get it ready for testing. Even though he didn’t admit it, the engineer seemed impressed with the result, if not the code.

By late afternoon, they had a functioning interactive prototype. Kate fed ChatGPT the prototype link and asked it to create a usability testing script. It came up with a basic, but complete test script, including a checklist for observers to take notes.

Kate went through the script carefully and updated it to add probing questions about AI transparency, emotional check-ins, more specific task scenarios, and a post-test debrief that looped back to the sprint goal.

Kate did a dry run with her product partner, who teased her: “Did you really need me? Couldn’t your AI do it?” It hadn’t occurred to her, but she was now curious!

“Act as a Gen Z user seeing this interactive prototype for the first time. How would you react to the language, steps, and tone? What would make you feel more confident or in control?”

It worked! ChatGPT simulated user feedback for the first screen and asked if she wanted it to continue. “Yes, please,” she typed. A few seconds later, she was reading what could have very well been a screen-by-screen transcript from a test.

Kate was still processing what she had seen as she drove home, happy she didn’t have to stay late. The simulated test using AI appeared impressive at first glance. But the more she thought about it, the more disturbing it became. The output didn’t mention what the simulated user clicked, and if she had asked, she probably would have received an answer. But how useful would that be? After almost missing her exit, she forced herself to think about eating a relaxed meal at home instead of her usual Prototype-Thursday-Multitasking-Pizza-Dinner.

Kate’s Reflection

Today was the most meta I’ve felt all week: building a prototype about AI, with AI, while being coached by AI. And it didn’t all go the way I expected.

While ChatGPT didn’t deliver prototype screens, Figma Make coded a working, interactive prototype with interactions I couldn’t have built in Figma Design. I used curiosity and experimentation today, by asking: What if I reworded this? What if I flipped that flow?

AI moved fast, but I had to keep steering. But I have to admit that tweaking the prototype by changing the words, not code, felt like magic!

Critical thinking isn’t optional anymore — it is table stakes.

My impromptu ask of ChatGPT to simulate a Gen Z user testing my flow? That part both impressed and unsettled me. I’m going to need time to process this. But that can wait until next week. Tomorrow, I test with 5 Gen Zs — real people.

Friday: Test (aka Prototype Meets User)

Day 5 in a design sprint is a culmination of the week’s work from understanding the problem, exploring solutions, choosing the best, and building a prototype. It’s when teams interview users and learn by watching them react to the prototype and seeing if it really matters to them.

As Kate prepped for the tests, she grounded herself in the sprint problem statement and the users: “How might we help Gen Z users confidently take their first financial action in our app — in a way that feels simple, safe, and puts them in control?”

She clicked through the prototype one last time — the link still worked! And just in case, she also had screenshots saved.

Kate moderated the five tests while her product and engineering partners observed. The prototype may have been AI-generated, but the reactions were human. She observed where people hesitated, what made them feel safe and in control. Based on the participant, she would pivot, go off-script, and ask clarifying questions, getting deeper insights.

After each session, she dropped the transcripts and their notes into ChatGPT, asking it to summarize that user’s feedback into pain points, positive signals, and any relevant quotes. At the end of the five rounds, Kate prompted them for recurring themes to use as input for their reflection and synthesis.

The trio combed through the results, with an eye out for any suspicious AI-generated results. They ran into one: “Users Trust AI”. Not one user mentioned or clicked the ‘Why this?’ link, but AI possibly assumed transparency features worked because they were available in the prototype.

They agreed that the prototype resonated with users, allowing all to easily set their financial goals, and identified a couple of opportunities for improvement: better explaining AI-generated plans and celebrating “win” moments after creating a plan. Both were fairly easy to address during their product build process.

That was a nice end to the week: another design sprint wrapped, and Kate’s first AI-augmented design sprint! She started Monday anxious about falling behind, overwhelmed by options. She closed Friday confident in a validated concept, grounded in real user needs, and empowered by tools she now knew how to steer.

Kate’s Reflection

Test driving my prototype with AI yesterday left me impressed and unsettled. But today’s tests with people reminded me why we test with real users, not proxies or people who interact with users, but actual end users. And GenAI is not the user. Five tests put my designerly skill of observation to the test.

GenAI helped summarize the test transcripts quickly but snuck in one last hallucination this week — about AI! With AI, don’t trust — always verify! Critical thinking is not going anywhere.

AI can move fast with words, but only people can use empathy to move beyond words to truly understand human emotions.

My next goal is to learn to talk to AI better, so I can get better results.

Conclusion

Over the course of five days, Kate explored how AI could fit into her UX work, not by reading articles or LinkedIn posts, but by doing. Through daily experiments, iterations, and missteps, she got comfortable with AI as a collaborator to support a design sprint. It accelerated every stage: synthesizing user feedback, generating divergent ideas, giving feedback, and even spinning up a working prototype, as shown below.

What was clear by Friday was that speed isn’t insight. While AI produced outputs fast, it was Kate’s designerly skills — curiosity, empathy, observation, visual communication, experimentation, and most importantly, critical thinking and a growth mindset — that turned data and patterns into meaningful insights. She stayed in the driver’s seat, verifying claims, adjusting prompts, and applying judgment where automation fell short.

She started the week on Monday, overwhelmed, her confidence dimmed by uncertainty and the noise of AI hype. She questioned her relevance in a rapidly shifting landscape. By Friday, she not only had a validated concept but had also reshaped her entire approach to design. She had evolved: from AI-curious to AI-confident, from reactive to proactive, from unsure to empowered. Her mindset had shifted: AI was no longer a threat or trend; it was like a smart intern she could direct, critique, and collaborate with. She didn’t just adapt to AI. She redefined what it meant to be a designer in the age of AI.

The experience raised deeper questions: How do we make sure AI-augmented outputs are not made up? How should we treat AI-generated user feedback? Where do ethics and human responsibility intersect?

Besides a validated solution to their design sprint problem, Kate had prototyped a new way of working as an AI-augmented designer.

The question now isn’t just “Should designers use AI?”. It’s “How do we work with AI responsibly, creatively, and consciously?”. That’s what the next article will explore: designing your interactions with AI using a repeatable framework.

Poll: If you could design your own AI assistant, what would it do?

• Assist with ideation?
• Research synthesis?
• Identify customer pain points?
• Or something else entirely?

Share your idea, and in the spirit of learning by doing, we’ll build one together from scratch in the third article of this series: Building your own CustomGPT.

Resources

• Sprint: How to Solve Big Problems and Test New Ideas in Just Five Days, by Jake Knapp
• The Design Sprint
• Figma Make
• “OpenAI Appeals ‘Sweeping, Unprecedented Order’ Requiring It Maintain All ChatGPT Logs”, Vanessa Taylor

Tools

As mentioned earlier, ChatGPT was the general-purpose LLM Kate leaned on, but you could swap it out for Claude, Gemini, Copilot, or other competitors and likely get similar results (or at least similarly weird surprises). Here are some alternate AI tools that might suit each sprint stage even better. Note that with dozens of new AI tools popping up every week, this list is far from exhaustive.

(yk)

Artificial intelligence is increasingly automating large parts of design and development workflows — tasks once reserved for skilled designers and developers. This streamlining can dramatically speed up project delivery. Even back in 2023, AI-assisted developers were found to complete tasks twice as fast as those without. And AI tools have advanced massively since then.

Yet this surge in capability raises a pressing dilemma:

Does the environmental toll of powering AI infrastructure eclipse the efficiency gains?

We can create websites faster that are optimized and more efficient to run, but the global consumption of energy by AI continues to climb.

As awareness grows around the digital sector’s hidden ecological footprint, web designers and businesses must grapple with this double-edged sword, weighing the grid-level impacts of AI against the cleaner, leaner code it can produce.

The Good: How AI Can Enhance Sustainability In Web Design

There’s no disputing that AI-driven automation has introduced higher speeds and efficiencies to many of the mundane aspects of web design. Tools that automatically generate responsive layouts, optimize image sizes, and refactor bloated scripts should free designers to focus on completing the creative side of design and development.

By some interpretations, these accelerated project timelines could represent a reduction in the required energy for development, and speedier production should mean less energy used.

Beyond automation, AI excels at identifying inefficiencies in code and design, as it can take a much more holistic view and assess things as a whole. Advanced algorithms can parse through stylesheets and JavaScript files to detect unused selectors or redundant logic, producing leaner, faster-loading pages. For example, AI-driven caching can increase cache hit rates by 15% by improving data availability and reducing latency. This means more user requests are served directly from the cache, reducing the need for data retrieval from the main server, which reduces energy expenditure.

AI tools can utilize next-generation image formats like AVIF or WebP, as they’re basically designed to be understood by AI and automation, and selectively compress assets based on content sensitivity. This slashes media payloads without perceptible quality loss, as the AI can use Generative Adversarial Networks (GANs) that can learn compact representations of data.

AI’s impact also brings sustainability benefits via user experience (UX). AI-driven personalization engines can dynamically serve only the content a visitor needs, which eliminates superfluous scripts or images that they don’t care about. This not only enhances perceived performance but reduces the number of server requests and data transferred, cutting downstream energy use in network infrastructure.

With the right prompts, generative AI can be an accessibility tool and ensure sites meet inclusive design standards by checking against accessibility standards, reducing the need for redesigns that can be costly in terms of time, money, and energy.

So, if you can take things in isolation, AI can and already acts as an important tool to make web design more efficient and sustainable. But do these gains outweigh the cost of the resources required in building and maintaining these tools?

The Bad: The Environmental Footprint Of AI Infrastructure

Yet the carbon savings engineered at the page level must be balanced against the prodigious resource demands of AI infrastructure. Large-scale AI hinges on data centers that already account for roughly 2% of global electricity consumption, a figure projected to swell as AI workloads grow.

The International Energy Agency warns that electricity consumption from data centers could more than double by 2030 due to the increasing demand for AI tools, reaching nearly the current consumption of Japan. Training state-of-the-art language models generates carbon emissions on par with hundreds of transatlantic flights, and inference workloads, serving billions of requests daily, can rival or exceed training emissions over a model’s lifetime.

Image generation tasks represent an even steeper energy hill to climb. Producing a single AI-generated image can consume energy equivalent to charging a smartphone.

As generative design and AI-based prototyping become more common in web development, the cumulative energy footprint of these operations can quickly undermine the carbon savings achieved through optimized code.

“

Water consumption forms another hidden cost. Data centers rely heavily on evaporative cooling systems that can draw between one and five million gallons of water per day, depending on size and location, placing stress on local supplies, especially in drought-prone regions. Studies estimate a single ChatGPT query may consume up to half a liter of water when accounting for direct cooling requirements, with broader AI use potentially demanding billions of liters annually by 2027.

Resource depletion and electronic waste are further concerns. High-performance components underpinning AI services, like GPUs, can have very small lifespans due to both wear and tear and being superseded by more powerful hardware. AI alone could add between 1.2 and 5 million metric tons of e-waste by 2030, due to the continuous demand for new hardware, amplifying one of the world’s fastest-growing waste streams.

Mining for the critical minerals in these devices often proceeds under unsustainable conditions due to a lack of regulations in many of the environments where rare metals can be sourced, and the resulting e-waste, rich in toxic metals like lead and mercury, poses another form of environmental damage if not properly recycled.

Compounding these physical impacts is a lack of transparency in corporate reporting. Energy and water consumption figures for AI workloads are often aggregated under general data center operations, which obscures the specific toll of AI training and inference among other operations.

And the energy consumption reporting of the data centres themselves has been found to have been obfuscated.

Reports estimate that the emissions of data centers are up to 662% higher than initially reported due to misaligned metrics, and ‘creative’ interpretations of what constitutes an emission. This makes it hard to grasp the true scale of AI’s environmental footprint, leaving designers and decision-makers unable to make informed, environmentally conscious decisions.

Do The Gains From AI Outweigh The Costs?

Some industry advocates argue that AI’s energy consumption isn’t as catastrophic as headlines suggest. Some groups have challenged ‘alarmist’ projections, claiming that AI’s current contribution of ‘just’ 0.02% of global energy consumption isn’t a cause for concern.

Proponents also highlight AI’s supposed environmental benefits. There are claims that AI could reduce economy-wide greenhouse gas emissions by 0.1% to 1.1% through efficiency improvements. Google reported that five AI-powered solutions removed 26 million metric tons of emissions in 2024. The optimistic view holds that AI’s capacity to optimize everything from energy grids to transportation systems will more than compensate for its data center demands.

However, recent scientific analysis reveals these arguments underestimate AI’s true impact. MIT found that data centers already consume 4.4% of all US electricity, with projections showing AI alone could use as much power as 22% of US households by 2028. Research indicates AI-specific electricity use could triple from current levels annually by 2028. Moreover, Harvard research revealed that data centers use electricity with 48% higher carbon intensity than the US average.

Advice For Sustainable AI Use In Web Design

Despite the environmental costs, AI’s use in business, particularly web design, isn’t going away anytime soon, with 70% of large businesses looking to increase their AI investments to increase efficiencies. AI’s immense impact on productivity means those not using it are likely to be left behind. This means that environmentally conscious businesses and designers must find the right balance between AI’s environmental cost and the efficiency gains it brings.

Make Sure You Have A Strong Foundation Of Sustainable Web Design Principles

Before you plug in any AI magic, start by making sure the bones of your site are sustainable. Lean web fundamentals, like system fonts instead of hefty custom files, minimal JavaScript, and judicious image use, can slash a page’s carbon footprint by stripping out redundancies that increase energy consumption. For instance, the global average web page emits about 0.8g of CO₂ per view, whereas sustainably crafted sites can see a roughly 70% reduction.

Once that lean baseline is in place, AI-driven optimizations (image format selection, code pruning, responsive layout generation) aren’t adding to bloat but building on efficiency, ensuring every joule spent on AI actually yields downstream energy savings in delivery and user experience.

Choosing The Right Tools And Vendors

In order to make sustainable tool choices, transparency and awareness are the first steps. Many AI vendors have pledged to work towards sustainability, but independent audits are necessary, along with clear, cohesive metrics. Standardized reporting on energy and water footprints will help us understand the true cost of AI tools, allowing for informed choices.

You can look for providers that publish detailed environmental reports and hold third-party renewable energy certifications. Many major providers now offer PUE (Power Usage Effectiveness) metrics alongside renewable energy matching to demonstrate real-world commitments to clean power.

When integrating AI into your build pipeline, choosing lightweight, specialized models for tasks like image compression or code linting can be more sustainable than full-scale generative engines. Task-specific tools often use considerably less energy than general AI models, as general models must process what task you want them to complete.

There are a variety of guides and collectives out there that can guide you on choosing the ‘green’ web hosts that are best for your business. When choosing AI-model vendors, you should look at options that prioritize ‘efficiency by design’: smaller, pruned models and edge-compute deployments can cut energy use by up to 50% compared to monolithic cloud-only models. They’re trained for specific tasks, so they don’t have to expend energy computing what the task is and how to go about it.

Using AI Tools Sustainably

Once you’ve chosen conscientious vendors, optimize how you actually use AI. You can take steps like batching non-urgent inference tasks to reduce idle GPU time, an approach shown to lower energy consumption overall compared to requesting ad-hoc, as you don’t have to keep running the GPU constantly, only when you need to use it.

Smarter prompts can also help make AI usage slightly more sustainable. Sam Altman of ChatGPT revealed early in 2025 that people’s propensity for saying ‘please’ and ‘thank you’ to LLMs is costing millions of dollars and wasting energy as the Generative AI has to deal with extra phrases to compute that aren’t relevant to its task. You need to ensure that your prompts are direct and to the point, and deliver the context required to complete the task to reduce the need to reprompt.

Additional Strategies To Balance AI’s Environmental Cost

On top of being responsible with your AI tool choice and usage, there are other steps you can take to offset the carbon cost of AI usage and enjoy the efficiency benefits it brings. Organizations can reduce their own emissions and use carbon offsetting to reduce their own carbon footprint as much as possible. Combined with the apparent sustainability benefits of AI use, this approach can help mitigate the harmful impacts of energy-hungry AI.

You can ensure that you’re using green server hosting (servers run on sustainable energy) for your own site and cloud needs beyond AI, and refine your content delivery network (CDN) to ensure your sites and apps are serving compressed, optimized assets from edge locations, cutting the distance data must travel, which should reduce the associated energy use.

Organizations and individuals, particularly those with thought leadership status, can be advocates pushing for transparent sustainability specifications. This involves both lobbying politicians and regulatory bodies to introduce and enforce sustainability standards and ensuring that other members of the public are kept aware of the environmental costs of AI use.

It’s only through collective action that we’re likely to see strict enforcement of both sustainable AI data centers and the standardization of emissions reporting.

Regardless, it remains a tricky path to walk, along the double-edged sword of AI’s use in web design.

Use AI too much, and you’re contributing to its massive carbon footprint. Use it too little, and you’re likely to be left behind by rivals that are able to work more efficiently and deliver projects much faster.

“

The best environmentally conscious designers and organizations can currently do is attempt to navigate it as best they can and stay informed on best practices.

Conclusion

We can’t dispute that AI use in web design delivers on its promise of agility, personalization, and resource savings at the page-level. Yet without a holistic view that accounts for the environmental demands of AI infrastructure, these gains risk being overshadowed by an expanding energy and water footprint.

Achieving the balance between enjoying AI’s efficiency gains and managing its carbon footprint requires transparency, targeted deployment, human oversight, and a steadfast commitment to core sustainable web practices.

(yk)