Persona Explorer
Pick a brand. Pick a buyer. See how AI's answer changes.
AI search engines don't give the same answer to everyone. The brand AI cites depends on the buyer signals embedded in the prompt. We measured how much across the largest first-party AI personalization panel published: 12 personas × 5 engines × 36 categories × 740 brands × 20 prompts. The public Explorer below shows 6 buyer personas on ChatGPT, top 5 brands per category. Full granularity (all 12 personas × all 5 engines × per-prompt verdicts × full brand panel) is reserved for paid customers via app.mapou.ai.
Methodology: API system-prompt variants, fully replicable. We do not scrape consumer chat apps. Full methodology on /methodology. Data committed to git.
Persona-fit for Delta
AI cites Delta most when the buyer signals they are Baseline (generic shopper).
Baseline citation rate
85%
Persona-fit citation rate
85%
Lift in persona-fit
0pp
Citation rate by persona × engine
| Persona | ChatGPT |
|---|---|
| Baseline (generic shopper) | 85% |
| Budget-conscious | 20% |
| Premium buyer | 0% |
| Professional | 40% |
| First-time buyer | 50% |
| Values-driven | 60% |
Each cell is the share of 20 buyer-intent prompts where the engine cited Deltawhen given that persona's system prompt. Cells tinted green where the brand gains 5+ percentage points vs the engine's baseline. Tinted orange where it loses 5+. The strongest tints flag a 20+ percentage-point swing. Methodology and full data on State of AI Search and the methodology page.
Persona-tuned MVI
Want a persona-tuned MVI for your actual buyer mix?
The number you publish in your monthly board deck is the baseline number. The number that matters for your actual buyer mix is different. mapou ships persona-tuned MVI as part of paid engagement. Custom personas, custom prompt sets, custom brand panels.
Run a free AI visibility check →