Giving Users A Voice Through Virtual Personas

Turn scattered user research into AI-powered personas that give anyone consolidated multi-perspective feedback from a single question.

• Paul Boag
• Dec 23, 2025
• 0 comments

Giving Users A Voice Through Virtual Personas

• 10 min read
• User Research,

UX,
Design,

• Share on Twitter, LinkedIn

About The Author

Paul is a leader in conversion rate optimisation and user experience design thinking. He has over 25 years experience working with clients such as Doctors …
More about
Paul ↬

*Weekly tips on front-end & UX.
Trusted by 182,000+ folks.*

Turn scattered user research into AI-powered personas that give anyone consolidated multi-perspective feedback from a single question.

In my previous article, I explored how AI can help us create functional personas more efficiently. We looked at building personas that focus on what users are trying to accomplish rather than demographic profiles that look good on posters but rarely change design decisions.

But creating personas is only half the battle. The bigger challenge is getting those insights into the hands of people who need them, at the moment they need them.

Every day, people across your organization make decisions that affect user experience. Product teams decide which features to prioritize. Marketing teams craft campaigns. Finance teams design invoicing processes. Customer support teams write response templates. All of these decisions shape how users experience your product or service.

And most of them happen without any input from actual users.

The Problem With How We Share User Research

You do the research. You create the personas. You write the reports. You give the presentations. You even make fancy infographics. And then what happens?

The research sits in a shared drive somewhere, slowly gathering digital dust. The personas get referenced in kickoff meetings and then forgotten. The reports get skimmed once and never opened again.

When a product manager is deciding whether to add a new feature, they probably do not dig through last year’s research repository. When the finance team is redesigning the invoice email, they almost certainly do not consult the user personas. They make their best guess and move on.

This is not a criticism of those teams. They are busy. They have deadlines. And honestly, even if they wanted to consult the research, they probably would not know where to find it or how to interpret it for their specific question.

The knowledge stays locked inside the heads of the UX team, who cannot possibly be present for every decision being made across the organization.

What If Users Could Actually Speak?

What if, instead of creating static documents that people need to find and interpret, we could give stakeholders a way to consult all of your user personas at once?

Imagine a marketing manager working on a new campaign. Instead of trying to remember what the personas said about messaging preferences, they could simply ask: “I’m thinking about leading with a discount offer in this email. What would our users think?”

And the AI, drawing on all your research data and personas, could respond with a consolidated view: how each persona would likely react, where they agree, where they differ, and a set of recommendations based on their collective perspectives. One question, synthesized insight across your entire user base.

You can question how personas will react to different scenarios based on the research available. (Large preview)

This is not science fiction. With AI, we can build exactly this kind of system. We can take all of that scattered research (the surveys, the interviews, the support tickets, the analytics, the personas themselves) and turn it into an interactive resource that anyone can query for multi-perspective feedback.

Building the User Research Repository

The foundation of this approach is a centralized repository of everything you know about your users. Think of it as a single source of truth that AI can access and draw from.

If you have been doing user research for any length of time, you probably have more data than you realize. It is just scattered across different tools and formats:

• Survey results sitting in your survey platform,
• Interview transcripts in Google Docs,
• Customer support tickets in your helpdesk system,
• Analytics data in various dashboards,
• Social media mentions and reviews,
• Old personas from previous projects,
• Usability test recordings and notes.

The first step is gathering all of this into one place. It does not need to be perfectly organized. AI is remarkably good at making sense of messy inputs.

If you are starting from scratch and do not have much existing research, you can use AI deep research tools to establish a baseline.

[Research with perplexity]

Online deep research with a tool like perplexity can be invaluable as a starting point for user research. (Large preview)

These tools can scan the web for discussions about your product category, competitor reviews, and common questions people ask. This gives you something to work with while you build out your primary research.

Creating Interactive Personas

Once you have your repository, the next step is creating personas that the AI can consult on behalf of stakeholders. This builds directly on the functional persona approach I outlined in my previous article, with one key difference: these personas become lenses through which the AI analyzes questions, not just reference documents.

The process works like this:

• Feed your research repository to an AI tool.
• Ask it to identify distinct user segments based on goals, tasks, and friction points.
• Have it generate detailed personas for each segment.
• Configure the AI to consult all personas when stakeholders ask questions, providing consolidated feedback.

Here is where this approach diverges significantly from traditional personas. Because the AI is the primary consumer of these persona documents, they do not need to be scannable or fit on a single page. Traditional personas are constrained by human readability: you have to distill everything down to bullet points and key quotes that someone can absorb at a glance. But AI has no such limitation.

This means your personas can be considerably more detailed. You can include lengthy behavioral observations, contradictory data points, and nuanced context that would never survive the editing process for a traditional persona poster. The AI can hold all of this complexity and draw on it when answering questions.

You can also create different lenses or perspectives within each persona, tailored to specific business functions. Your “Weekend Warrior” persona might have a marketing lens (messaging preferences, channel habits, campaign responses), a product lens (feature priorities, usability patterns, upgrade triggers), and a support lens (common questions, frustration points, resolution preferences). When a marketing manager asks a question, the AI draws on the marketing-relevant information. When a product manager asks, it pulls from the product lens. Same persona, different depth depending on who is asking.

Personas can have different lenses relevant to different functions within the business. (Large preview)

The personas should still include all the functional elements we discussed before: goals and tasks, questions and objections, pain points, touchpoints, and service gaps. But now these elements become the basis for how the AI evaluates questions from each persona’s perspective, synthesizing their views into actionable recommendations.

Implementation Options

[...]

Original source