Source: Articles on Smashing Magazine — For Web Designers And Developers | Read More
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. Paul Boag shows how to breathe new life into this stale UX asset and demonstrates that it’s possible to create truly useful functional personas in a lightweight way.
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.
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.
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.
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:
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.
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:
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.
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:
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.
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:
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.
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.
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.
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:
Save the report you get back into your project.
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.
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:
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.
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.
As you work through the above process, you will encounter problems. Here are common pitfalls and how to avoid them:
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:
With this approach, personas evolve from static deliverables into dynamic reference points your whole team can rely on.
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.
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.