What does a GEO report look like?
A worked SAMPLE report for a made-up business, section by section, and where each number would come from on a real account.
This page shows a SAMPLE weekly GEO report for a fictional business, Example Dental Co. Every number in it is made up. A useful report answers five questions: what your AI-readiness score is and what moved, which customer questions name you on which engines, what AI gets wrong about you, which rivals get named instead, and what to fix next.
What should a weekly GEO report tell you?
GEO (Generative Engine Optimization) is the work of getting a business found, trusted and quoted by AI search. A report on it is only useful if it leads to a decision. Five questions do that:
- Can AI engines read my site, and did that change? An AI-readiness score, tracked week to week.
- Do AI answers name me for the questions my customers ask? Per question, per engine.
- Is AI saying anything wrong about me? Hours, prices, services, locations.
- Who gets named instead of me? Rivals, and how often.
- What should I fix next? One ranked list, worst first.
Each section of the sample below maps to a feature our platform actually runs.
What does a sample weekly report look like?
1. Headline (SAMPLE)
| Measure | This week | Last week |
|---|---|---|
| AI-readiness score | 70 / 100 | 60 / 100 |
| AI crawlers getting in | 40 of 42 | 38 of 42 |
| Wrong facts found | 2 | 3 |
| Open fixes | 5 (1 marked do first) | 7 |
| Rivals named instead of you | 3 of 10 tracked | 3 of 10 tracked |
2. Tracked questions, by engine (SAMPLE)
| Question | ChatGPT | Perplexity | Google AI Overviews | Gemini | Copilot |
|---|---|---|---|---|---|
| best family dentist in Exampleton | Cited, #2 | Cited, #1 | Not cited | Cited, #3 | Not cited |
| emergency dentist open Saturday Exampleton | Not cited | Cited, #2 | Not cited | Not cited | Cited, #1 |
| is Example Dental Co any good | Cited, #1 | Cited, #1 | Cited, #1 | Cited, #1 | Cited, #1 |
A real run covers 12 engines; the sample shows 5 to fit the page.
3. Wrong facts (SAMPLE)
| Engine | What it said | Your verified fact | Severity | Fix |
|---|---|---|---|---|
| Engine A | "Closed on Saturdays" | Open Saturday, 9 to 1 | High | Correction drafted |
| Engine B | "Founded in 1901" | Founded in 2001 | Medium | Correction drafted |
4. Rivals (SAMPLE)
| Brand | Share of voice |
|---|---|
| Rival A | 30% |
| Example Dental Co. | 20% |
| Rival B | 20% |
| Rival C | 10% |
| Nobody named yet | 20% |
5. Next fixes, worst first (SAMPLE)
| Priority | Fix | Why |
|---|---|---|
| Do first | Allow the two blocked AI crawlers in robots.txt | A blocked crawler can't read the site at all |
| 2 | Publish Saturday hours on the contact page and in structured data | Fixes the high-severity wrong fact at the source |
| 3 | Add a page answering "emergency dentist on Saturday" | The question where most engines skip you |
Where does each number come from on a real account?
| Report section | Platform feature | How it is measured |
|---|---|---|
| AI-readiness score | AI Access | 30+ checks on every page we can find, rolled into one 0–100 score. A re-scan confirms each fix. |
| Crawlers getting in | AI Access / crawler test | 42 AI crawlers are tested for real, not just read from robots.txt. |
| Tracked questions | Prompt Simulator | Each saved question runs on 12 engines. Every answer is labelled live search, model memory or voice reconstruction, with your citation and rank read from the answer itself. |
| Wrong facts | Hallucination Watch | Up to 16 checkable questions per run, judged against your verified facts. Opinions are never flagged. Each wrong answer gets a drafted correction. |
| Rivals | Competitors | Share of voice counted from real AI answers, for up to 25 rivals you name, including the share nobody holds yet. |
| Next fixes | Fix list | Every finding from every scan in one list, ranked worst first, with steps. |
If you serve local customers, Local adds map-pack position on real searches, listings checked on 13 platforms, and name, address and phone conflicts on your own pages.
How should I read a report like this?
- Watch the trend, not the week. AI answers move. OpenAI itself warns that search results and citations "can be incomplete, outdated, or incorrect." OpenAI Help Center Four weeks of the same questions tell you more than one good Tuesday.
- Treat "not cited" as useful. An honest miss tells you which question to work on. A report that never shows a miss isn't measuring much.
- Keep "unknown" apart from "no". If an engine failed to answer, that is missing data, not a lost citation.
- Fix at the source. A wrong fact is fixed by publishing the correct one where engines read, not by emailing the AI company. See AI is wrong about your business.
- Don't expect a special file to fix Google. Google says no special AI files or markup are needed for its AI features; the pages have to be indexable and worth quoting. Google Search Central
For the metrics behind a report, see how to measure AI visibility.
How do I get a report like this for my business?
Start with the free audit, which shows where you stand before you pay for anything. Every self-serve plan includes all of the features these numbers come from: AI Pulse at $99 a month, AI Pro at $299 and AI Max at $999. They differ only in tokens, seats and markets. See pricing.
Sources
- ChatGPT search — OpenAI Help Center. Read Sep 24, 2026.
- AI features and your website — Google Search Central. Read Sep 24, 2026.
Common questions
Are the numbers on this page real?
No. Example Dental Co. is a fictional business and every figure in the sample is made up to show the layout.
How often should a GEO report run?
Weekly is a sensible rhythm for tracked questions.
Can one report cover several locations?
Our plans count markets: 1 on AI Pulse, 3 on AI Pro and 10 on AI Max. Ask us how your locations would map.
Does a good report mean guaranteed AI placement?
No. Nobody controls another company's AI engine. A report shows what moved against a real baseline, so you can see whether the work is paying off.
Related pages.
See what AI says about your business.
Enter your domain and the free AI visibility check shows where ChatGPT, Gemini, Perplexity and Google's AI answers mention you — and where they don't.