Free AEO Report: What Goes In It, How To Build One Free, And How The Models Behind The Answers Score

A free AEO report is a snapshot of how AI answer engines like ChatGPT, Perplexity, Gemini and Google's AI features name, describe and cite your brand, and you can build one yourself for free.

AEO stands for answer engine optimisation, which is the work of getting your brand named and cited when people ask AI tools a question.

The problem it solves is that your normal SEO reports show rankings and clicks, but they never show you the written answer a buyer reads or the competitors named in it.

If you'd like my team to look at your AI visibility live, you can book a free strategy session and get a custom AI visibility plan.

In this post I'll explain what goes in the report, how to build it with three AI prompts, which free sources help and how to read the results.

Near the end, I'll show how the AI models behind those answers score on this leaderboard, and what that does and doesn't tell you.

What is a free AEO report?

A free AEO report is a short document that records what AI tools say when real buyers ask about your category or your brand.

Google says traffic from AI Overviews and AI Mode is counted inside the "Web" search type of the Search Console Performance report, so you can't see it on its own there.

The AEO report pulls that hidden layer into the open by looking at the answers themselves.

Nobody sells you the raw material, because it's just answers you collect and organise.

What goes in a free AEO report?

These are the nine sections I'd include, in the order a busy reader needs them.

I run every prompt at least three times, because AI answers change from one chat to the next.

How to build a free AEO report with three AI prompts

Start with a one-page fact sheet listing your name, services, locations, proof and the facts people often get wrong.

Then use the first prompt to draft your buyer questions, and cut anything a real customer wouldn't type.

I am building an AEO report for [BRAND], a [CATEGORY] business serving [CUSTOMER] in [LOCATION].
List 30 questions a real buyer would type into ChatGPT, Perplexity, Gemini or Google before choosing a company like mine.
Group them as best-of, comparison, problem and brand-name questions, and keep my brand out of the first three groups.

Ask each question three times in ChatGPT, Perplexity, Gemini and Google, use a fresh chat each time and save every answer with its source links and the date.

Then use the second prompt to turn the saved answers into a table.

Read the AI answers below, saved from [AI TOOL] on [DATE].
Return one table row per answer with: prompt, run, [BRAND] named (yes or no), competitors named, URLs cited, facts stated about [BRAND], tone.
Use only what the answers say, and write "not stated" when something is missing.

Check five random rows by hand, because a model can still misread a long answer.

Finally, use the third prompt to turn the table into the report.

Using the table below and my brand fact sheet, calculate share of answers for us and each competitor, list the ten most cited domains,
mark every stated fact as correct, wrong or out of date, and write a one-page summary with three fixes for the next 30 days.
Use only numbers from the table and do not promise any results.

Re-check the percentages in a spreadsheet, because language models are better at reading than counting.

Don't skip the crawl check, which takes two minutes in a browser.

OpenAI's documentation says sites that opt out of OAI-SearchBot won't be shown in ChatGPT search answers, and Perplexity recommends allowing PerplexityBot if you want to appear in its results.

Google says a page must be indexed and eligible to show with a snippet to appear as a supporting link in AI Overviews or AI Mode.

📞 Want a strategist to read your report with you? Book a free strategy session with Goldie Agency and leave with a custom visibility plan, a clear AI visibility roadmap and a competitor gap analysis.

Which free sources help?

I checked these on their own pages on 17 October 2026.

HubSpot's grader is a reputation view, while your own prompt runs show live answers, so the two make a useful pair.

How to read a free AEO report

Fix crawl access first, because a blocked crawler makes every other section pointless.

Fix wrong facts second, because the cited URL in your accuracy table usually points at the page or profile that needs changing.

Then study the citation sources, because they show where the AI learns about your category.

Goldie Agency's strategy session page says AI platforms don't recommend brands simply because those brands publish more on their own websites.

It says they look for consistent signals across the wider web, such as trusted publications, expert articles, independent reviews and comparisons.

That's why the list of outside sites citing your competitors is often the most useful line in the report.

How often should you run a free AEO report?

I'd run the first report by hand, then repeat it once a month with exactly the same prompts.

Monthly is often enough to see a trend, and it's rare enough that each report still feels worth reading.

Run an extra report after any big change, such as a site migration, a rebrand or a new product launch.

Keep every old report, because the comparison between months is worth more than any single snapshot.

Common mistakes with a free AEO report

The first mistake is testing only prompts that include your brand name, which makes almost any business look good.

The second mistake is running each prompt once and treating the answer as fact.

The third mistake is letting the AI do the maths without checking it in a spreadsheet.

The fourth mistake is changing the prompt set every month, which makes the trend impossible to read.

The fifth mistake is building a beautiful report and never fixing the first item on it.

What the free strategy session is, and isn't

The free strategy session with Goldie Agency is a live call with an experienced strategist, and the page says it's 100% free with no obligation.

You leave with a custom visibility plan, an AI visibility roadmap, a competitor gap analysis and a plan for traffic, mentions and leads.

It isn't a written or PDF report, so the best move is to build your own report first and bring it along.

What a free AEO report can't tell you

A free AEO report is a snapshot, so it can't predict what an AI tool will say next week.

It can't explain exactly why a model picked one source, because no AI company publishes that formula.

It can't guarantee anything, and nobody can promise what a model will say.

What it gives you is an honest baseline, so next month's report shows whether your work moved anything.

📞 Not sure which gap to fix first? Grab a free strategy session and see why competing brands appear more often, plus where your business can build a stronger presence.

Which AI models sit behind the answers, and how do they score on GoldieBench?

Now for the GoldieBench part, and I need to be clear about what it measures.

GoldieBench scores models on one-shot builds such as landing pages, apps, games and visuals, on a 0 to 10 scale.

It doesn't measure which brands an AI recommends, and it doesn't measure how accurate an AI is about your business.

A free AEO report doesn't ship its own model, and neither do the free tools above.

The assistants you test are built by model makers, and these are how those makers' models score on their live pages today.

Claude Opus 5 is the highest of these at 8.27, with GPT-5.6 Sol close behind at 8.16.

Gemini 3.6 Flash averages 7.08, and it's the closest view I have of the Gemini family on the board.

Perplexity and Microsoft Copilot have no model on GoldieBench, so I can't give you a score for either.

The board also can't tell you which exact model a consumer app runs at any moment, because the makers change that themselves.

So please don't read these scores as a ranking of which assistant is easiest to appear in.

Which model should tabulate your AEO report?

The scores matter more for the second half of the job, which is reading your saved answers and filling in the table.

That work rewards a model that follows a detailed instruction exactly, because the table has to match the source text line for line.

A model that follows a brief well first time usually makes fewer mistakes for you to catch.

Even so, the checks around the model matter more than the model itself.

Asking the model to compare its own table with the raw answers, and doing the maths in a spreadsheet, will catch more errors than switching models.

Any of the top models above can do this job well, as long as you never let it invent a number.

I'd also keep the same model for every monthly report, because switching models can change how borderline answers get labelled.

If you do switch, re-run last month's table with the new model first so you know whether a change came from the answers or from the model reading them.

Related reading on GoldieBench

My King of AEO post looks at answer engine optimisation from the model side.

My AI SEO optimization agency post explains the work that usually follows a report like this.

In the same batch, I wrote about the AEO audit behind the report and the free GEO audit checks from the original GEO research.

If you'd like expert eyes on your results, book a free strategy session and leave with a plan for traffic, mentions and leads.

Also On Our Network

• Agent Operatorsthe seven-step operator build of a free AEO report
• AI Income Deskfour business use cases for a free AEO report
• AI Tool Verdictevery AEO report section and free source, rated
• agentos.guidea one-hour free AEO report quick start for your Agent OS
• agentos.guidethe Claude AI SEO system I use to rank pages

Build it once by hand, re-run the same prompts every month and let the numbers guide your fixes, because that's how a free AEO report earns its keep.