Synthetic, Qualitative, and "Vector" Personas: What's the Difference (and When to Use Each)?
We are observing three distinct types of personas emerging across teams, but not all personas are created equal.
Recently, Arizona State actually pitted human-developed personas (qualitative) against AI-generated personas (synthetic), and when judged, the human personas "won." However, they found that combining the two (vector) could be the best option. Let's take a look.
1) Qualitative research-based personas
Most of us have been creating personas for decades, using a tried-and-tested process, grounded in interviews, field studies, and observed behavior. They're slower to build and create, but are great at driving genuine empathy and durable decision-making because they reflect actual user motives, constraints, and context. Great for building empathy within teams.
Best for: product strategy, prioritization, and aligning cross-functional teams.
Watch out for: can go stale; they require a maintenance cadence.
2) Synthetic (AI-generated) personas
AI-constructed "users" produced from models and secondary data. In seconds, they can simulate needs, objections, and scenarios, which can be helpful for hypothesis generation and concept exploration; however, they can also hallucinate and introduce bias if not anchored to real evidence. Treat them as strawmen, not truth.
Best for: rapid exploration, "what-ifs," seeding test ideas.
Watchouts: origin, bias, hallucinations, and overconfidence—validate with real users before decisions.
3) "Vector personas" (the hybrid)
Vector personas are research-grounded personas that are supplemented with AI. You utilize vetted artifacts, such as interviews, observational notes, tickets, feedback, and analytics, but then supplement them with expanded details that AI can provide.
Best for: teams that want qualitative truth but with additional AI details.
Watchouts: governance, version changes, and require source citations in outputs.
The article "Leveraging AI Toward the Development of Vector Personas for UX Research" goes into great detail about the differences and benefits of vector personas. One great example was filling in on vague participant information: "In contrast to [the AI's] detailed response, our realtor interviewees responded to this more technical question with vague answers. One of our realtors responded, "Honestly, I love data, and the more the better for me. So, I'd be interested in anything.""
A checklist for using AI in persona work.
Provenance: every AI-generated assertion must cite a human-collected source (doc/link/time).
Bias & drift checks: schedule reviews; compare persona statements to fresh interviews and analytics.
Change log: version personas; highlight what changed and why (new data? segment shift?).
Decision policy: synthetic = ideation only; qualitative/vector = decision-grade when evidence passes a threshold.
Need a quick way to explain?
Qualitative personas anchor teams to reality.
Synthetic personas accelerate exploration—but need human guardrails.
Vector personas blend the two, giving you speed and explainability by grounding AI with fundamental research.
While human responses are often messy, they're real; AI, on the other hand, usually sounds too idealized and textbook. However, with terabytes of text, LLMs have details that most humans would find difficult to recall, but they can also try to please to the point of making incorrect inferences or conclusions. It sounds like the superpower is combining the two to create a more detailed view of a person, role, or group, thereby creating a fuller "vector" persona that utilizes the best of both worlds. 👩🦰 🤝 🤖
