Back to library

Synthetic Persona Builder

When you can't get real users in the room — tight timeline, niche segment, sensitive context, or early-stage concept — synthetic personas give you a structured way to surface blind spots before you invest in real research. This is not a replacement for talking to users. It's a tool for finding the questions worth asking when you finally do.

---

Context

What synthetic personas are good for:
  • Exposing assumptions you didn't know you were making
  • Finding obvious gaps in a feature concept
  • Generating a richer interview guide
  • Rapid internal alignment
  • What they are not good for:
  • Validating whether users will actually use a feature
  • Understanding emotional responses, trust, or behaviour change
  • Replacing even a single real user conversation
  • ---

    Step 1 — Define the persona set parameters

    Ask: product/feature, user segment (specific), number of personas (recommend 3), decision to inform, and dimensions of difference.

    Step 2 — Build each persona

    For each: specific role and company context, relationship to the product area (current behaviour, frustration, workaround, goal), AI adoption profile, and technical confidence level. Every detail should connect to a product implication.

    Step 3 — Run the persona against the product/feature

    Four responses per persona: First encounter (what they notice, assume, and ask), Objection (friction, objection, abandonment trigger), Ideal outcome (problem solved, colleague recommendation), and Edge case (unusual but realistic scenario).

    Step 4 — Cross-persona comparison

    Table across all personas: first reaction, primary objection, biggest value, most likely failure point, willingness to pay.

    Step 5 — Extract product implications

    Three outputs: Assumptions surfaced (what was assumed + which persona challenged it), Questions for real research (what to ask when you get to real users), and Feature implications (decisions suggested, with Low confidence label requiring validation).

    Quality check before delivering

    Each persona is grounded in specific, realistic context
    Every detail connects to a product implication
    All four response types are run for each persona
    Comparison table covers all personas
    Product implications distinguish synthetic insight from validated finding
    Limitation statement is present at the top
    Suggested next step: Use the "Questions for real research" section to build your next interview guide with aipm-user-interview. Synthetic personas are most valuable when they drive you toward better real research.