AI search visibility trackera citation sampling log
See how often you are named, and how sure that is.
Try the sampling logThe film is unavailable. You can read what it shows below.
What the film shows, in words
- How often are you named, and how sure is that?
- 8 answers logged. 3 named the brand.
- 38% observed. It could be as low as 14%, or as high as 69%.
- If a second run came out the same: still 38%. 18% to 61%.
- Five runs: 24% to 53%.
- Observed in this sample only. Not a universal score.
The same 38%. A very different certainty.
ExampleStep 1 of 4An invented sample · Illustrative example · No assistant is queried
Eight answers. Three named the brand.
95% intervalIt could be as low as 14%
Observed 38% 3 of 8
or as high as 69%
8 prompts logged · 2 competitor cited instead
- Cited with a linkLink
- Mentioned, no linkMention
- A competitor cited insteadCompetitor
- Not mentionedNone
- Dashed outline: no date recordedNo date
Mention rate by engine (cited or mentioned)
- ChatGPT1 of 4 · 95% interval 5% to 70%
- Perplexity2 of 4 · 95% interval 15% to 85%
The faintest setting is the lowest rate eight answers allow, the strongest the highest. Between them stands what was observed.
- 1 prompt has no date. Answers change over time; undated observations cannot be reproduced.
- Some prompts have no location or language. Answers vary by place and language; record both.
- ChatGPT: only 4 prompts. The 95% interval (5% to 70%) is wide. Run more prompts before drawing conclusions.
- Perplexity: only 4 prompts. The 95% interval (15% to 85%) is wide. Run more prompts before drawing conclusions.
Observed in this sample only. It is not a universal AI visibility score.
Now your own sample. Log what you saw.
Log each prompt you ran, with the engine, date and outcome. It does not run prompts for you. The rows below are an invented sample: replace them with what you saw.
Mention rate by engine (cited or mentioned)
- ChatGPT1 of 4 · 95% interval 5% to 70%
- Perplexity2 of 4 · 95% interval 15% to 85%
Read each line with its interval. A few answers on one engine say little about that engine.
- ChatGPT1 of 4 · 95% interval 5% to 70%25%
- Perplexity2 of 4 · 95% interval 15% to 85%50%
| Prompt | Engine | Date | Location | Outcome |
|---|---|---|---|---|
| Who does skills audits in Ghana? | ChatGPT | 20 Sept 2026 | Accra, English | Mentioned, no link |
| Best firms for programme evaluation in West Africa | ChatGPT | 20 Sept 2026 | Accra, English | A competitor cited instead |
| How to run a training needs assessment | ChatGPT | 20 Sept 2026 | Accra, English | Not mentioned |
| Who does skills audits in Ghana? | Perplexity | 21 Sept 2026 | Accra, English | Cited with a link |
| Best firms for programme evaluation in West Africa | Perplexity | 21 Sept 2026 | Accra, English | A competitor cited instead |
| How to run a training needs assessment | Perplexity | 21 Sept 2026 | Accra, English | Cited with a link |
| Ghana data protection registration help | Perplexity | 21 Sept 2026 | – | Not mentioned |
| AI governance consultants Africa | ChatGPT | Undated | London, English | Not mentioned |
Method checks
- Gap: 1 prompt has no date.Answers change over time; undated observations cannot be reproduced.
- Check: Some prompts have no location or language.Answers vary by place and language; record both.
- Note: ChatGPT: only 4 prompts.The 95% interval (5% to 70%) is wide. Run more prompts before drawing conclusions.
- Note: Perplexity: only 4 prompts.The 95% interval (15% to 85%) is wide. Run more prompts before drawing conclusions.
Observed in this sample only. It is not a universal AI visibility score.
XLSX Sample log
| Prompt | Engine | Date | Outcome |
|---|---|---|---|
| Who does skills audits in Ghana? | ChatGPT | 20 Sept 2026 | Mentioned, no link |
| Best firms for programme evaluation in West Africa | ChatGPT | 20 Sept 2026 | A competitor cited instead |
| How to run a training needs assessment | ChatGPT | 20 Sept 2026 | Not mentioned |
| Who does skills audits in Ghana? | Perplexity | 21 Sept 2026 | Cited with a link |
| Best firms for programme evaluation in West Africa | Perplexity | 21 Sept 2026 | A competitor cited instead |
and 3 more rows
GKM Strategic Advisory & AI
Citation sample log and the reproducible method
Your organisation
95% intervalIt could be as low as 14%
Observed 38% 3 of 8
or as high as 69%
8 prompts logged · 2 competitor cited instead
Observed in this sample only. It is not a universal AI visibility score.
PDF · 24 September 2026
Your full report
Citation sample log and the reproducible method, by email once delivery is connected.
Citation sample log and the reproducible method (spreadsheet + PDF); not a universal AI score
Tool: AI citation sampling Brand: Example Advisory Result: 38% mentioned (n = 8) Request: please scope AI search visibility work (SVC-077).Send as a brief
How the rate is worked out, and what it cannot say.
The method is published so you can check it. Every figure on this page comes from it.
Method
- 01
Mention rate = prompts where you were cited or mentioned ÷ prompts run, per engine and overall.
- 02
Uncertainty is the Wilson 95% interval, which stays sensible for small samples.
- 03
Every observation needs a date, engine and location to be reproducible.
What it cannot tell you
- A sample is not the whole picture; answers vary by wording, user, place and time.
- It does not run prompts for you.
- No universal score is given, because none exists.
Quality ruleRecord engine/date/location and uncertainty; cap runtime costs.
The log says what you saw. What to do about it is a separate question.
A sample records what a handful of answers said on the days you asked. Deciding which questions matter, and what your own pages say in reply to them, is work this log does not do.
-
Kenneth GagakumaTool ownerBook Kenneth · from USD 90