Product-Market Fit: Building Rigorous Personas with Generative AI

A few years ago, I joined a fast-moving project in which product and design decisions were being made weekly around the digital experience, including what features to build next, how to align with real customer needs, and how to keep up with constant releases. At the same time, the marketing team was working in parallel on copy, imagery, and messaging—often without a clear understanding of who the end user was.

We all agreed we needed a better understanding of our users. But tight timelines and limited resources across UX teams kept research deprioritized. As a result, many key design and content decisions were made based on internal assumptions of what we believed made sense at the time.

That experience stuck with me: It’s not unique to one team.

As researchers and designers, we know the cost of skipping foundational UX work. And yet, it still happens. Why? We didn’t lack user data; we lacked the time and structure to synthesize it into something meaningful and actionable. UX expert Bill Albert, PhD, points out that evaluating product usefulness and desirability at the concept stage, including core components of product-market fit, should be within the scope of UX. Yet one of the most overlooked and undervalued artifacts in this process is personas.

When created thoughtfully, personas capture the essence of your target audience—not just their demographics—but their goals, frustrations, and motivations. Personas serve as a shared reference point that helps cross-functional teams align product decisions, anchor conversations, and clarify the value that the product is truly delivering.

Personas: Still Useful, Still Misunderstood

A good persona helps answer a fundamental question: Who are we designing for and what matters to them?

It helps align the team around real user goals, expectations, and pain points. It also helps prioritize features by giving product and engineering teams a relatable anchor for decision-making.

As Aurora Harley of the Nielsen Norman Group describes it, personas are “fictional characters” based on user research that help designers understand and communicate about the people they’re designing for.

But personas take time and expertise to develop. In some teams, they’re seen as static, oversimplified documents that don’t evolve with the product. In others, they’re skipped altogether due to limited resources. And in many teams, creating structured, shareable insights—like user personas—can feel like a luxury rather than a necessity.

Worse, some teams try to shortcut the process with AI tools that auto-generate personas with no real data backing them. These tools often rely on generic demographic assumptions and give no transparency into how insights are derived.

That’s not helpful. But it doesn’t mean AI has no place in the process.

Figure 1: An image of professionals brainstorming ways to use AI (designed by Freepik™).

Using Generative AI in Your UX Process

Here’s the opportunity: Generative AI gives us a creative way to reduce that tension. With the right prompts and process, AI can help us synthesize qualitative data faster, build draft personas more efficiently, and bring user insights to life without compromising the integrity of our work.

Prompt engineering is the practice of crafting precise inputs that guide AI toward useful outputs. It’s both an art and a science, focused on understanding how large language models interpret prompts and how to iterate for better results.

Used thoughtfully, generative AI can amplify the research process, not replace it. It can help us:

analyze large volumes of qualitative data quickly,
spot recurring themes or user motivations, and
draft personas that reflect real users based on real conversations.

Read more about prompt engineering at LearnPrompting.org.

Tips for Incorporating Generative AI in Your UX Research

Shift Your Mindset: Treat AI like a Collaborator

Using generative AI effectively means shifting your mindset. As AI expert Jeremy Utley, adjunct professor of AI at Stanford, puts it, “Think of AI like a teammate, and say ‘what do you need to know from me to get the best response?’” Instead of treating it like a tool that gives you instant answers, treat it like a collaborator, that is, another researcher or design partner.

That mindset is essential for getting high-quality, relevant, and safe outputs from any AI system.

Be Specific and Structured

Generic prompts yield generic answers. Be clear about your role (“You are a UX researcher…”), the task (“summarize this transcript…”), and the format you want the output in (“into bullet points, quotes, or relevant document types.”). Specificity reduces ambiguity and gives you more actionable responses.

Read more about prompt engineering at LearnPrompting.org.

Use Iterative Prompting

Treat AI interactions like a conversation. Start with a broad prompt, then refine your request based on the response. For example, after a summary, you can follow up with, “Group these responses into common themes,” or “Identify behavioral patterns across these transcripts.”

Read more about iterative prompting at Google Cloud™ AI Guide.

A Practical Workflow: Creating Personas with Generative AI

Start with Research Planning

Generative AI can support this stage if you clearly define the criteria and goals you’re working with. Start by creating a focused research plan:

Define your research objectives.
Prepare recruitment and screening materials.
Draft your interview or discussion guide.

As part of this plan, always check with your legal or compliance team to ensure privacy protocols are followed before storing or sharing any transcripts.

Figure 2: A list of interview questions suggested by ChatGPT after I inputted research objectives and criteria for discussions with users.

Capture and Transcribe User Conversations

Conduct your interviews and record them (with consent). Then transcribe them using transcription features in tools like Otter™, Descript™, or Zoom™.

Explore Initial Themes and Patterns with Your AI Prompts

Once you’ve reviewed your transcripts, you can begin working with generative AI to explore themes. Here’s a sample prompt structure: “You’re helping analyze user interviews. I’ll provide transcripts one at a time. Summarize each, then identify patterns and one to three representative personas at the end.”

This process ensures you’re guiding the AI, not letting it guess blindly.

Figure 3: This is an example of the themes identified, which ChatGPT suggested after I provided a test transcript from an interview from an IBM® course.

Review and Refine the Output

Don’t just copy and paste the AI-generated personas into your deck. Instead, do this:

Compare the output with your own notes.
Check for hallucinations or inaccuracies.
Fill in gaps with contextual knowledge.

AI might spot patterns faster than you can, but it doesn’t know what’s important to your research unless you tell it.

Evangelize Insights Across Your Team

Once you’ve refined your personas, share them with your team, going beyond the slides. Consider doing the following:

Create quick videos or Looms™ to explain each persona.
Map user goals to product features in a shared doc.
Annotate real user quotes to build empathy.

Pro tip: You might even use your persona as an assistant to help answer questions from product teams. As mentioned earlier, when treating AI as a partner, Jeremy Utley suggests: “Construct a psychological profile of your conversation partner—the large language model—to roleplay a conversation and give you feedback from their perspective.” In this case, you can treat the persona as the customer and talk to it directly. The goal isn’t just to create personas; it’s to make them usable and bring them to life!

Use AI Responsibly and Transparently

We can’t talk about AI without talking about responsibility. When using generative tools, keep these ethical considerations in mind:

Be transparent with participants about how their data will be used.
Avoid over-relying on AI for synthesis: It’s there to support your thinking, not replace it.
Disclose in your research documentation that generative tools were part of the process.
Consult your legal and data privacy team if you’re handling sensitive information.

Organizations like the AI Now Institute and UNESCO have issued strong guidance on responsible and transparent AI practices. This is something every practitioner should be familiar with.

Generative AI Isn’t a Magic Wand but a Powerful Ally

What excites me most about generative AI isn’t that it can replace parts of UX work; it’s that it can make it more accessible.

Just imagine this:

Teams can build draft personas faster with generative AI’s help from planning, to during, and after the research, reducing development from weeks to days.
Researchers can spend more time validating insights instead of sorting transcripts.
Stakeholders can be part of the process of witnessing and participating in your research activities. Because it’s distilled, they can more easily interact with the persona outcomes.

Generative AI isn’t perfect, but neither are rushed product decisions made without user insight. If we use it with intention, transparency, and creativity, we can move faster without leaving the user behind.

Final Thoughts

Whether you’re a solo UX practitioner at a startup or a researcher supporting multiple squads, you don’t have to choose between speed and rigor.

Tools like ChatGPT, Claude™, and other LLMs give us a new lens to look at data, a new way to craft narratives, and a new bridge between what users need and what teams can build.

So the next time someone says, “We don’t have time for personas,” try this approach. You might prove that—with the right AI assistance—you don’t have time not to.

User Experience

Product-Market Fit: Building Rigorous Personas with Generative AI

Personas: Still Useful, Still Misunderstood

Using Generative AI in Your UX Process

Tips for Incorporating Generative AI in Your UX Research

A Practical Workflow: Creating Personas with Generative AI

Generative AI Isn’t a Magic Wand but a Powerful Ally

Final Thoughts

Further Reading

Siyun Hur

About UXPA

USER EXPERIENCE MAGAZINE

UXPA

FOLLOW