Free to use, keep and change. Works in Excel and Google Sheets. In Google Sheets, use File, then Import, then Upload. If Excel opens it in Protected View, click Enable Editing so the formulas calculate.
What it does
The tracker turns the method in how to track your AI visibility into a file you can use today. You write down the questions your customers ask, ask them in each AI tool once a month, and log what comes back. The workbook does the counting.
What is inside
| How to use | The method, and a legend for which cells to fill in |
|---|---|
| Questions | Space for 30 questions, each with an ID and a type: category, comparison, problem, local or brand |
| Log | One row per question per tool: whether you were named, your position, competitors named, sources cited, and anything wrong about you. Tool names and yes or no answers are dropdowns, so the counts stay clean |
| Summary | For each tool and month, how many logged answers named you, written as "3 of 25", plus a count of answers that got something wrong |
It covers ChatGPT, Claude, Perplexity, Gemini, Microsoft Copilot, Google AI Overviews and Google AI Mode. Use the ones your customers use and ignore the rest.
What it takes to do properly
The spreadsheet costs nothing. Doing the tracking properly takes real time. Every question needs a fresh chat, so earlier conversations do not sway the answer. It has to happen on the same day each month, or the months stop being comparable. And every answer has to be read closely enough to note who was named and which sources it relied on.
| Questions | 25 |
|---|---|
| AI tools | 7 |
| Answers to read and log | 175 |
| Time per answer | About two minutes, as a rough estimate |
| Time each month | Close to six hours, before any follow-up |
One tool takes under an hour. All the tools your customers use, every month, without skipping one, adds up to most of a working day.
Logging is the easy part
The spreadsheet shows where you stand. Improving it is separate work, and it needs time and experience most businesses would rather spend elsewhere:
- Correcting what AI tools get wrong about you, at the listings and pages they took it from.
- Getting described well where it counts: the directories, review sites and publications that keep appearing under Sources cited.
- Rewriting your own pages so they answer your customers' questions plainly enough to be quoted.
- Reading the trend, and telling a real change from the normal month-to-month variation in AI answers.
The full version: what we run for clients
This tracker is a basic version of our process. It covers one part of it: asking the questions and logging the answers. The full version, run every month, adds the data a spreadsheet cannot collect on its own and the work that changes the answers.
| Free tracker | Clear Mountain, every month | |
|---|---|---|
| Asking the questions | You run it | We run it in every tool your customers use |
| The question list | You write it | Written with you, in your customers' words, and kept fixed so months compare |
| Visits from AI tools | Not included | Tracked in your analytics: visits from ChatGPT, Perplexity, Copilot and Gemini |
| Copilot citations | Not included | Tracked in Bing Webmaster Tools' AI Performance report |
| Google traffic and searches for your name | Not included | Tracked in Search Console |
| Fixing what AI tools get wrong | Up to you | Done for you |
| Work on the sites AI tools cite | Up to you | Done for you |
| Reporting | Your spreadsheet | A plain-English monthly summary, every number with its source |
How we measure each of these, and what none of them can show: how we measure AI visibility.
What it cannot tell you
Whether you run it or we do, the Summary shows counts on purpose, not percentages. AI answers vary between users and sessions, so what gets logged is a sample of your own questions. It shows direction over several months. It does not show your share of every AI answer in your category, and nothing you can buy does either without estimating.