# AI Discovery sample report

> Public example reviewed 10 September 2026. The values below are illustrative and are not a customer measurement.

See how Prompeteer separates website readiness, answer evidence, citations, recommendation position, sentiment and action planning in one dated report.

## Readiness snapshot

- **Website readiness: 84 / 100.** Public pages were reachable and the key offer was parseable in this illustration.
- **Brand mentions: 8 / 24.** Eight of twenty-four completed answers named the sample business.
- **Citations: 14 / 24.** Fourteen completed answers included at least one recorded source.
- **Recommendation position: Top three in 5 / 8.** Five of eight eligible answers placed the sample business in the first three recommendations.
- **Brand sentiment: 62% positive.** Eight named answers were assessed: five positive, two neutral and one mixed.
- **AEM actions: 4 proposed.** Four evidence-linked actions are ready for owner review; proposal is not implementation proof.

## Evidence coverage by section

| Report section | Coverage | Recorded evidence |
| --- | --- | --- |
| Website readiness | Observed | **84 / 100.** Seven checks passed, one needs review and two were not applicable. |
| Answer collection | Measured | **24 / 24 complete.** Three web-grounded engines and five model-memory proxies completed the same six questions. |
| Brand mentions | Measured | **8 / 24 named.** A mention rate is calculated from completed answers and is separate from citation rate. |
| Citation evidence | Measured | **14 / 24 answers cited.** Recorded citations include source URLs and exact destination pages when available. |
| Recommendation position | Measured | **5 / 8 top-three.** Position is reported only for answers that contain an explicit ordered recommendation. |
| Brand sentiment | Measured | **8 assessed mentions.** The classifier labels the answer wording and stores a quote for every assessed mention. |
| Forecast outlook | Not yet measured | **History required.** A forecast is withheld until comparable dated reports have enough history for validation. |

## Model and layer results

| Model | Layer | Completion | Named | Citations |
| --- | --- | --- | --- | --- |
| OpenAI API | Web-grounded | 5 / 6 complete | 2 named | 3 citations |
| Perplexity API | Web-grounded | 6 / 6 complete | 3 named | 5 citations |
| Gemini API | Web-grounded | 6 / 6 complete | 2 named | 4 citations |
| Meta API | Web-grounded | 6 / 6 complete | 1 named | 2 citations |
| Grok API | Web-grounded | 6 / 6 complete | 2 named | 3 citations |
| DeepSeek API | Model memory | 6 / 6 complete | 0 named | — |
| Qwen API | Model memory | 6 / 6 complete | 0 named | — |
| GLM API | Model memory | 5 / 6 complete | 0 named | — |

The same six buyer questions run across each configured engine. A zero named rate is a measured result, not a failed call. Web-grounded answers contribute citation evidence; model-memory answers do not.

## Method and next step

Prompeteer asks stable, buyer-relevant questions, records raw answers and citations, extracts explicit recommendations and stores exact quotes for sentiment review. Comparisons use compatible dates and question sets. The private report adds the answer matrix, selected competitor context, prior versions and an AEM work queue.

[Read the measurement guide](https://prompeteer.ai/guides/ai-visibility-measurement) · [See the AEM workflow](https://prompeteer.ai/aem) · [Choose a plan](https://prompeteer.ai/pricing)

Canonical: https://prompeteer.ai/aeo/sample-report