Search Personas

The Persona page answers a question your metrics can't: who is asking the questions you're tracked in? Maya clusters the behavior in your evaluation data — the phrasing, intents, and journeys behind your prompts and answers — into search personas, so your content targets people, not keywords.

Generating personas

Click Generate Personas and Maya builds a set from your accumulated evaluation data. You need some run history first — the empty state tells you to collect data before generating. Personas can be deleted and regenerated at any time, and individual personas can be edited inline if you want to correct a name or detail.

What a persona contains

Each card describes a distinct searcher type:

• Experience level — Beginner, Intermediate, Advanced, or Expert.
• Awareness stage — Unaware → Problem Aware → Solution Aware → Product Aware → Most Aware, the classic funnel position.
• Confidence — high/medium/low, how strongly the data supports this persona.
• Goals and pain points / frustrations — what they're trying to do and what blocks them.

How to use them

1. Map prompts to personas. A "Beginner / Problem Aware" persona asks discovery questions ("how do I even solve X"); a "Product Aware" one asks comparisons. Check your prompt set covers each persona you care about.
2. Write to the stage. Content for Problem Aware readers explains the problem before the product; Most Aware readers need offers and proof, not education. Assistants mirror this — answers to beginner questions cite explainer content, not product pages.
3. Spot missing personas. If an important customer type produced no persona, your tracked prompts don't represent them — add prompts in their voice.

Notes

• Personas describe AI-search behavior observed in your data, not your CRM segments. Expect overlap but not identity.
• Low-confidence personas are hypotheses; collect more runs before investing content against them.