Short answer: a free monthly SEO report template in Excel that puts Google Search Console numbers next to AI answer numbers. It shows how often AI answers name your brand and cite your site, puts a 95% range on each rate, and says whether a change between two periods is outside normal run-to-run noise. Fill in the shaded cells; the rest is calculated for you.
Free template · Excel · CC BY 4.0
SEO and AI visibility report template
Five tabs: Report, Google, Pages, AI answers and Start here. Fictional example rows included. No signup.
What is inside
| Tab | What it does | What you fill in |
|---|---|---|
| Report | One page: the latest Search Console month, then two tables (answers that name you, answers that cite you) comparing two periods by engine, each rate with its range and a verdict. | The two period labels to compare |
| CTR and month-on-month change in clicks and impressions. | Clicks, impressions and average position for each month | |
| Pages | Click change and CTR for your top pages, this month against last. | Page, clicks and impressions for both months |
| AI answers | Named rate and cited rate per period and engine, each with a 95% Wilson interval. Flags rows with under 30 answers. | Answers logged, answers naming you, answers citing your site |
| Start here | Instructions and method notes. | Nothing |
How to fill it in
- Google tab. From the Performance report in Search Console, enter clicks, impressions and average position for each month, oldest first, with the month written as 2026-10.
- Pages tab. Enter your top pages with clicks and impressions for this month and last month.
- AI answers tab. One row per period and engine. Ask each of your buyer questions at least three times per engine on each check date, then count the answers, the answers that name your brand, and the answers that cite a page on your site. If you use the free answer log analyzer, add up the runs, named and cited columns of its cells CSV for your brand.
- Report tab. Type the two periods you want to compare in the shaded cells, written exactly as on the AI answers tab. Engine names in the shaded cells must match the AI answers tab too.
Worked example
The example rows are the totals from the answer log analyzer’s own worked example: a fictional Dublin accounting firm, 12 questions, three runs each on three engines, 108 answers in September and 108 in October. So you can check the template against the analyzer’s published report.
| Engine | Named, September | Named, October | Change | Verdict |
|---|---|---|---|---|
| All engines | 24% (17% to 33%) | 33% (25% to 43%) | +9.3 points | Within noise (p = 0.13, ranges overlap) |
| ChatGPT | 28% (16% to 44%) | 33% (20% to 50%) | +5.6 points | Within noise |
| Copilot | 14% (6% to 29%) | 19% (10% to 35%) | +5.6 points | Within noise |
| Perplexity | 31% (18% to 47%) | 47% (32% to 63%) | +16.7 points | Within noise |
A nearly 17-point rise on one engine looks like a result worth reporting to a client. At 36 answers per month, the sheet says it is not distinguishable from noise. That is the conversation this template is built to have before someone asks why the next month went back down.
Why the verdict is strict
The sheet says “Change outside noise” only when two things are true: a two-proportion test gives p below 0.05, and the two 95% ranges do not overlap. The second condition is there because repeated answers to the same question are not independent: an engine asked the same question again tends to give a similar answer, so the test on its own overstates how sure you can be. Rows with fewer than 30 answers in either period are marked “Too few answers to compare” and get no verdict, even though the p-value and overlap columns still show.
Read many rows at once with care. Across a page of comparisons, about one in twenty will cross p below 0.05 by chance when nothing changed. The full method is in the analyzer’s methods file.
What it will not tell you
- AI Overview clicks on their own. Google includes AI Overviews and AI Mode traffic in the Web search type of the Performance report (Google, AI features and your website), and that page does not describe a way to filter it out. For a query-level view, use the AI Overview CTR loss finder.
- Why a rate moved. Engine model updates and retrieval changes land between check dates too, and the sheet cannot separate them from your own work. Change one thing at a time if you want to know what worked.
- Anything about questions you did not ask. The rates describe your question set, nothing wider. Keep the questions fixed so periods compare.
Related
- GEO checklist, graded by evidence: what to change once the report shows a gap.
- AI Overview CTR loss finder: which queries lost clicks to AI answers.
- Question set builder: the fixed questions this report depends on.
- Keyword research for AI search: how to find the questions buyers ask.
- How long does it take a new website to get indexed and ranked in Google?: our day-by-day timing on a brand-new domain, useful for setting expectations before the first report.
- All templates.
License and credit
Free to use and adapt under CC BY 4.0, including for client work, with credit and a link: AI Visibility Census, “SEO report template with AI visibility”, aivisibilitycensus.com/templates/seo-report-template/. The file that builds the sheet is on GitHub. Version 1 is archived on Zenodo with a permanent DOI: 10.5281/zenodo.23960473.
New templates and every Census test result, sent when they publish: get the Census by email. Free, and one click to leave.
Update log
| Date | Change |
|---|---|
| 4 November 2026 | Template published (version 1). |
| 4 November 2026 | Version 1 archived on Zenodo (DOI 10.5281/zenodo.23960473). |