What a monthly GEO report should include

By: Irina Shvaya | October 13, 2026
This article is part of our GEO and AEO Guide: how AI search chooses sources, how to get cited, and how to budget for and choose help. Want to know where your own site stands? Request a free GEO audit.

We suggest eight parts for a useful monthly GEO report: a short summary with decisions, Google AI impressions, Microsoft AI citations, visits from AI assistants, organic search context, results from a fixed set of test prompts, leads and sales, and the work done with next month's plan. Each part should name the source of its numbers, and most of those sources are free reports from Google and Microsoft.

GEO (generative engine optimization) is the work of getting your business mentioned and cited in AI-generated answers. Reporting on it is harder than reporting on ads, because no single dashboard covers every assistant, and for several assistants we could not confirm any report for site owners. A good report is honest about that. It shows what can be measured, labels what is an estimate, and says what can't be seen.

This guide works whether you write the report yourself, get it from an agency or receive it from a freelancer. Use it as a checklist against whatever lands in your inbox each month.

Key Takeaways

  • Every number in a GEO report should name its source and its date range. A figure with no source is not evidence.
  • Three free first-party sources carry most of the weight: Search Console's Generative AI performance report, Bing Webmaster Tools' AI Performance report and Google Analytics 4's AI Assistant channel.
  • Google's report counts impressions, Microsoft's counts citations and GA4 counts visits. They measure different things and should never be added together.
  • A vendor's "AI visibility score" is that vendor's estimate. Google states: "No third-party tool has access to our internal ranking or AI systems."
  • Microsoft says its AI Performance data "represents a sample of overall citation activity," so read it as a trend and not a full count.
  • The report should end with decisions: what was done, what happens next and what needs your approval.

What the report is for

A report exists to help you decide something: keep going, change direction, or stop. If a 30-page PDF arrives each month and nobody acts differently because of it, it isn't working, however good the charts look.

Three tests help. Can you tell in two minutes whether things improved? Can you trace each number to a source you could log in to and check? Does it say what will happen next month and why? If the answer to any of these is no, ask for changes.

The eight sections and where each number comes from

SectionSourceWhat it showsWhat it can't show
1. Summary and decisionsWritten by a personWhat changed, why it matters, what is needed from youNot applicable
2. Google AI impressionsSearch Console, Generative AI performance reportHow often links to your pages were shown in AI Overviews and AI ModeClicks, queries, the answer text
3. Microsoft AI citationsBing Webmaster Tools, AI PerformanceCitations across Copilot and Bing's AI summaries, and grounding queriesClicks or traffic; a complete count
4. Visits from AI assistantsGoogle Analytics 4, AI Assistant channelSessions that arrived with a recognized AI referrer, and what those visitors didGoogle's AI features; visits with no referrer
5. Organic search contextSearch Console, classic Performance reportClicks, impressions, CTR and average position; brand searchesAI clicks separated from other clicks
6. Prompt set resultsManual tests or a tracking toolWhether you are mentioned, linked and described accurately for fixed questionsHow many real people ask those questions
7. Leads and salesYour CRM, forms and call logsBusiness results, including what customers say brought them inA clean split between SEO and GEO effects
8. Work done and next stepsThe team doing the workWhat changed on the site and off it, and the planNot applicable

Section by section

1. Summary and decisions

Half a page, written in plain language. It should say what moved, the likely reason, anything that went wrong, and the decisions waiting on you. If the summary only restates numbers from the pages behind it, it has skipped the hard part.

2. Google AI impressions

Google's Generative AI performance report counts "how many times links to your site were shown to a user in a generative AI feature on Google Search." It covers AI Overviews and AI Mode, and you can split it by page, country, date and device. The report should show the month's total, the comparison with the previous month, and the pages that gained or lost most.

It should not call these clicks or rankings. This report has neither. We go through it in detail in how to read the Generative AI performance report. If your site doesn't have the report yet, the section should say so. Google's stated reason is usually too few impressions in AI features.

3. Microsoft AI citations

Bing Webmaster Tools has an AI Performance report, announced as a public preview on February 10, 2026, covering Microsoft Copilot, AI-generated summaries in Bing and what Microsoft calls select partner integrations. The metrics to include are Total Citations, Average Cited Pages, and the grounding queries your pages were cited for.

Two of Microsoft's own caveats belong in the report beside the numbers. On its help page, Microsoft says "The data shown represents a sample of overall citation activity." It also answers the question of whether citations represent clicks or traffic with a plain no: a citation "does not represent traffic, clicks, or user engagement." Setup and the full metric list are in our guide to Bing Webmaster Tools and IndexNow for AI search.

4. Visits from AI assistants

GA4's default channel group has an AI Assistant channel. Google's definition says it covers visits "from sources like ChatGPT, Gemini, Deepseek, Copilot, or Grok" and "excludes Google's AI Overviews and AI Mode." The section should show sessions, the pages those visitors landed on, and the conversions they completed.

The label matters. This is "AI-referred visits we can identify," not total AI traffic. Visits without a referrer are processed as Direct, and Google doesn't publish the list of assistants it matches. The AI Assistant channel in GA4 explains the limits.

5. Organic search context

AI visibility doesn't sit apart from regular search. Google says that optimizing for its generative AI features is "still SEO," and AI impressions are a subset of the web totals in the classic Performance report. So the report needs the classic numbers too: clicks, impressions, CTR and average position, plus the top pages and queries.

Brand search belongs here. Search Console has a branded and non-branded queries filter, which Google says is not available for sites with a low number of impressions and may misidentify some queries. A steady rise in people searching for you by name is consistent with growing awareness. It doesn't prove where the awareness came from.

6. Prompt set results

The platform reports say nothing about what an answer said. A prompt set fills that gap: a fixed list of customer questions, run on a schedule, with the answers logged. The report should show how many runs mentioned you, how many linked to you, any inaccurate statements, and which other businesses were named, broken out by question group and by assistant.

Check two things. The prompt list should be the same as last month, with any additions flagged. And the counts should be stated with their sample, such as the number of runs behind each figure. How to build a prompt set covers the method.

7. Leads and sales

This section connects visibility to money, and it draws on your systems, not the vendor's. Include leads and sales from organic and AI-referred visits, and the answers to a "How did you hear about us?" question on forms and calls. Self-reported answers are imperfect, but they catch the buyer who heard your name from an assistant and arrived another way.

If the vendor has no access to your CRM, the section will be thin. That is a setup problem worth fixing early, since it is the part of the report a finance team cares about.

8. Work done and next steps

A dated list of what was changed: pages rewritten, facts corrected, listings updated, technical fixes shipped, coverage earned. Then next month's plan, with a reason for each item drawn from the data in the earlier sections.

Numbers to treat with care

  • A single "AI visibility score." If a tracking tool gives you one combined score, it can be a handy trend line inside that tool. It is still the vendor's estimate, built on the vendor's prompts and method, and it can't be compared across tools. Ask what goes into it. Google's guidance on optimizing for generative AI features is direct: "No third-party tool has access to our internal ranking or AI systems."
  • "Rankings" in an assistant. Answers vary between runs and users. A claim that you "rank #1 in ChatGPT" describes one answer on one day.
  • Totals added across platforms. Impressions, citations and sessions are different units from different systems.
  • "Estimated AI traffic" or "estimated value." If a figure is modeled, the report should say how. As of October 2026, we could not confirm any site-owner report from OpenAI, Perplexity or Anthropic that would supply such numbers.

Why the numbers won't match each other

Someone will notice that Search Console and GA4 disagree. Some difference is normal. Google's explanation of performance data lists reasons including privacy, processing, time lag and time zones, and notes that tools such as Google Analytics "track traffic only from users who have enabled JavaScript in their browser." Search Console's performance dates are in Pacific Time.

Timing matters for a monthly report too. Google says collected data is usually available in two to three days, and the newest points in the AI report can be preliminary. A report pulled on the first of the month is working with unfinished data. A few days' delay gives cleaner numbers.

A template you can copy

Paste this into a spreadsheet and fill one row per metric. The rows are a starting point, not a standard.

MetricSourceThis monthLast monthNote
Google AI impressionsSearch Console, Generative AI performance reportChart total; state the search type
Top pages by AI impressionsSame report, Pages tabList the top few and the biggest movers
Total CitationsBing Webmaster Tools, AI PerformanceA sample, per Microsoft
AI Assistant sessionsGA4, Traffic acquisitionIdentified visits only
Conversions from AI Assistant sessionsGA4Name the events counted
Organic clicks and impressionsSearch Console, classic Performance reportIncludes AI feature data
Branded clicksSearch Console, branded filterIf the filter is available
Prompt runs with a mentionPrompt logState runs and assistants
Inaccurate statements foundPrompt logList each one
Leads and sales, with stated sourceCRM and formsInclude self-reported answers

For layout ideas beyond a spreadsheet, see GEO dashboards and reporting templates. For choosing which of these to treat as your headline measures, see new KPIs for GEO campaigns.

What to settle before the first report

  1. Access. The Search Console, Bing Webmaster Tools and GA4 accounts should belong to your business, with the vendor added as a user.
  2. Baseline. Record every metric in the template before work starts.
  3. Prompt list. Agree on the questions and freeze them.
  4. Delivery date and format. A set day each month, and a short call to go through it if you want one.
  5. Raw data. Ask for exports along with the summary, so you keep the history if you change vendors.

It helps to have reporting duties written into the agreement. GEO contract terms explained describes what that language usually covers.

Conclusion

A monthly GEO report earns its place when it names its sources, keeps impressions, citations and visits apart, labels estimates as estimates, and ends with decisions. Most of what it needs is free and first-party. The rest is a disciplined prompt log and your own lead data. If you want an outside view of where your site stands before you set a baseline, eSEOspace offers a free GEO audit of your website structure and visibility. More guides are in our GEO and AEO Guide.

Frequently asked questions

Is there one dashboard that shows all AI visibility?

No. Google reports on its own AI features, Microsoft on Copilot and Bing, and GA4 on visits it can identify. As of October 2026, we could not confirm an equivalent site-owner report from OpenAI, Perplexity or Anthropic. A monthly report stitches these together.

Should I trust an AI visibility score from a tool?

Treat it as that tool's estimate. It can be useful for watching a trend within one tool, as long as you know which prompts and assistants it is built from. It is not a platform metric.

What if my site has no Generative AI performance report?

Google says that usually means the site hasn't received enough impressions in its AI features. The report should note that, and lean on the other sections until the data appears.

Can a GEO report show return on investment?

It can show the inputs: identified AI visits, the leads they produced and what customers say brought them in. It can't attribute every sale, because some AI exposure never produces a trackable click. Say that plainly in the report.

Put this into action with eSEOspace

We help businesses grow with seo, geo & aeo that actually performs. Explore the services behind this guide:

Book a free strategy call →

Get a FREE Audit

We'll perform a comprehensive SEO, AEO, GEO & CRO audit of your website — completely free — and show you exactly how to outrank your competitors.

Don't have a site yet? Get in touch →

Get a FREE GEO/AEO/SEO Audit

We'll analyze your site's SEO, GEO, AEO & CRO — completely free — and show you exactly how to get found across Google and AI answers.

Don't have a site yet? Get in touch →

You Might Also like to Read