Disclosure: we are Fortitude Media and we sell a competing depth of AI visibility tracking, so read our take with that in mind.

Last checked: 10 June 2026. All product facts on this page were verified against public pricing pages and announcements on that date. Read our full methodology.

The short answer

The HubSpot AEO Grader gave you a free snapshot of how three AI engines (ChatGPT, Perplexity and Gemini) see your brand, scored across sentiment, presence quality, brand recognition, share of voice and market position. It did not look at Claude or Google AI Overviews, it did not test the buying questions specific to your category, and it did not tell you who is winning the recommendations instead of you or what to fix first. The plan below takes you from score to action in five steps, four of which cost nothing and need no tools at all.

What the grader told you

The AEO Grader gave you a structured first look at your AI visibility, and that is worth more than most free tools manage. HubSpot launched it at its Spring 2026 Spotlight as a one-off snapshot, no account needed, scoring five dimensions:

  • Sentiment: whether AI engines describe you positively, neutrally or negatively.
  • Presence quality: how substantial and accurate the mentions of you are.
  • Brand recognition: whether the engines know who you are at all.
  • Share of voice: how much of the conversation you occupy relative to others.
  • Market position: where the engines place you in your market.

If your scores were strong, that is a real signal. If they were weak, that is also a real signal. Either way, you now have a vocabulary for the problem, which is exactly what a free grader should give you.

What the grader could not tell you

The grader is a snapshot of three engines, so there are six things it could not show you, and they matter in roughly this order.

  1. Claude. The grader covers ChatGPT, Perplexity and Gemini. Claude is not in HubSpot's tracking claims, and B2B buyers researching considered purchases use it.
  2. Google AI Overviews. Also absent from the tracking claims, despite sitting on top of the search results a large share of your buyers still start from.
  3. Your category's buying prompts. A general snapshot cannot test the specific, intent-loaded questions your buyers ask, such as "best [your category] for UK mid-market firms".
  4. Named competitors. A share-of-voice score tells you how much room you occupy. It does not tell you which rival is being recommended in your place, question by question.
  5. History. A one-off snapshot has no trend line. You cannot tell whether you are rising, falling or flat. (HubSpot's paid tracking, $50 per month after a 28-day trial, starts the line on the day you subscribe.)
  6. What to fix first. A score describes the problem. It does not sequence the work.

None of that is a criticism of a free tool. It is the boundary of one, and the five steps below take you past it.

Step 1: audit your presence across all five engines yourself

Ask each engine the questions your buyers ask, and write down what comes back. This costs nothing but an hour. Open ChatGPT, Claude, Gemini, Perplexity and Google (for AI Overviews) and put the same five to ten questions to each: who they would recommend in your category, who the alternatives to your best-known competitor are, what they know about your company. Record three things per answer: are you mentioned, what is said about you, and who is named instead of you. Our guide on how to track which AI engines recommend your business gives you a worksheet for exactly this exercise.

The most common finding: a business looks fine in the grader's three engines and is invisible in Claude or misdescribed in Google AI Overviews. You want to know that this week, not next year.

Step 2: check your site's AI readability basics

Make sure the engines can actually read and understand your site before you spend a penny on anything else. Many weak AI visibility scores trace back to the website, not the brand: pages that never say plainly what the company does, who it serves and where it operates; key facts buried in images or scripts; no clear answers to the questions buyers ask. Work through our technical checklist for an AI-ready website and fix the basics first. This is unglamorous work, it is free if you do it yourself, and it raises the ceiling on everything that follows.

Step 3: identify your category's buying prompts

List the questions a real buyer would ask an AI engine in the week before they shortlist suppliers like you. Not "what is [your company]" but the prompts with money behind them: "best [category] provider for [industry]", "[competitor] alternatives UK", "how much does [category] cost", "[category] for companies with [specific need]". Aim for fifteen to twenty-five prompts. Ask your sales team what questions prospects arrive with; those questions were asked somewhere first, and increasingly that somewhere is an AI engine. This list becomes your measuring stick: any tool, agency or internal effort should be judged on how those specific prompts answer, not on a generic score.

Step 4: benchmark two competitors

Run your prompt list for your two most dangerous competitors and compare the answers side by side. Pick the rival who beats you in deals and the rival who is loudest in your market, and put the same prompts to the same five engines for each of them. You are looking for patterns: which engines favour them, what sources the answers cite, what claims about them get repeated. This usually reveals why they are recommended (a well-cited comparison page, a directory listing, a body of third-party coverage) and therefore what your own gap actually is. It turns "our score is low" into "they are winning these eight prompts because of these specific sources", which is something you can act on.

Step 5: decide whether you need a tool, help, or both

Choose based on who will do the work, because that is the real decision. After steps 1 to 4 you will be in one of three positions:

  • You found no real problem. Keep the free options running as a smoke alarm: HubSpot's grader periodically, or a free ongoing tracker, and get back to your business.
  • You found a problem and have the team to fix it. Buy monitoring at the depth you need and brief your team. Paid tracking runs from $50 per month (HubSpot) up through deeper platforms; our round-up of AI visibility tools for UK B2B compares them, and our guide to tools, agencies or both covers the choice in full.
  • You found a problem and nobody internally owns it. This is where a tool alone will not save you, because dashboards do not write content or earn coverage.

For the second and third positions, this is where our ladder exists, and we will describe it plainly rather than dress it up. Start with our Free AI Visibility Check, which covers all five engines including the two the grader skips; it is the natural second opinion on your grader score. Sentinel Free tracks one brand on five prompts with a monthly refresh, free. Paid Sentinel tiers (£199, £499 and £999 per month) add configurable prompt depth, named-competitor share of voice and up to twelve months of your category's history on day one. And if nobody on your team can do the work, Forge, our managed service from £3,950 per month, does the content, digital PR and website work, measured on the same chart. If HubSpot's $50 tracking is the right depth for your stage, take it; the steps above cost you nothing either way.

Frequently asked questions

What does the HubSpot AEO Grader actually measure?

It gives a free, one-off snapshot of how ChatGPT, Perplexity and Gemini see your brand, scored across five dimensions: sentiment, presence quality, brand recognition, share of voice and market position. No account is needed. It does not cover Claude or Google AI Overviews and does not track your category's specific buying prompts over time.

How do I improve my AEO score?

Start with the basics: make sure each AI engine can read your site and understands plainly what you do, fix the gaps our technical checklist covers, then build content and third-party coverage that answers your category's buying prompts better than your competitors do. Scores move when the sources the engines cite start saying the right things about you, which is content and PR work, not a settings change.

Is one AEO Grader score enough to act on?

No, treat it as a prompt for a fuller audit rather than a diagnosis. It covers three engines with a general snapshot, so before spending money, check all five engines yourself, test your category's real buying prompts and benchmark two competitors. The pattern across those checks tells you what to fix first.

Do I need a paid tool after the AEO Grader?

Only if you found a problem and need to track progress while fixing it. If your audit came back clean, free options are enough for now. If you found gaps, paid monitoring runs from $50 per month (HubSpot) to deeper platforms with more engines, competitor depth and history, and managed services exist for teams with nobody to do the work.

What is the difference between the AEO Grader and Fortitude's Free AI Visibility Check?

The grader is HubSpot's free snapshot of three engines: ChatGPT, Perplexity and Gemini. Our Free AI Visibility Check covers five engines: ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews. Running both costs nothing and shows you whether the engines the grader skips are costing you recommendations.

The free second opinion

If you only do one thing after reading this, do step 1. And if you would like it done for you, our Free AI Visibility Check covers all five engines, including Claude and Google AI Overviews, in one report. No card, no sales call, and you will know within minutes whether the grader showed you the whole picture.