HubSpot AEO Grader: What It Measures and Misses
HubSpot's AEO Grader, now called the AI Search Grader, is a free, one-time check that asks ChatGPT, Perplexity and Gemini how they describe your brand based on their training data, and scores the result out of 100 across five dimensions. It's a quick, useful read on how AI models perceive your brand, but as a single snapshot of model knowledge it can't tell you whether assistants recommend you when buyers ask for options, which sources they cite, or how results change over time.
This post describes what the tool measures, from HubSpot's own page as checked on 15 November 2026, and what a one-off grade like this leaves out. It isn't a criticism of HubSpot's tool, which is free and does what it says. It's a guide to reading the result well.
What the HubSpot AEO Grader is
The address many people know, hubspot.com/aeo-grader, now redirects to HubSpot's AI Search Grader. HubSpot's page still describes it in AEO terms, and most people still search for it as the HubSpot AEO Grader, so we use both names here.
According to HubSpot's page:
- It's free and one-off. HubSpot describes it as a free, one-time check, with no HubSpot account required. You complete a short form to see the full results.
- You enter four details: company name, location, products or services, and industry. It accepts any brand, so you can check a competitor too.
- It asks three AI platforms. HubSpot's FAQ says it currently analyses GPT-5.4 mini from OpenAI, Perplexity, and Gemini from Google, and asks each how it characterises your brand based on its training data.
- It's quick. HubSpot says results are ready in under two minutes.
What the HubSpot AEO Grader measures
The score out of 100 is made up of five dimensions, with the weights HubSpot publishes:
- Sentiment (up to 40 points). The overall tone AI uses about your brand, how that varies by topic, and the credibility of the sources shaping it.
- Presence quality (up to 20 points). How substantively AI discusses your brand, the authority of the sources behind that, and how complete the information is.
- Brand recognition (up to 20 points). How widely recognised your brand is across AI training data.
- Share of voice (up to 10 points). Where your brand stands against competitors in AI responses at the time of the check.
- Market competition (up to 10 points). How AI positions you: as a leader, challenger or niche player, and as an innovator, disruptor or traditionalist.
HubSpot also says the scoring is deterministic, with schema validation and retries, and that you get a written interpretation of the results. Sentiment carries the most weight, which HubSpot explains as measuring how AI characterises your brand, not just whether it recognises you.
What HubSpot AEO adds
HubSpot AEO is a separate, paid product. HubSpot's page describes it as ongoing monitoring that tracks visibility across specific prompts, compares competitors and gives prioritised recommendations, week over week, across ChatGPT, Perplexity and Gemini. As of November 2026 the page lists it at USD 50 a month, or included in Marketing Hub Professional and Enterprise. If you already use HubSpot, that's the obvious next step HubSpot offers after the grader.
What a one-off grade can't tell you
Every tool makes trade-offs, and a free two-minute check has to. Here's what's outside what the grader measures, by its own description.
Whether assistants recommend you when buyers ask
The grader asks AI platforms about your brand. That tells you how they describe you when your name is in the question. But buyers usually don't start with your name. They ask "which suppliers stock this?" or "what's the best tool for that?" Whether you're named in reply to those category questions is a different measurement, and it's the one closest to winning business.
What assistants find when they search
HubSpot says the grader works from the models' training data. Many real answers now involve a live web search: OpenAI's help page says ChatGPT may search the web automatically when a question would benefit from current information. A brand that's new, or that has changed recently, may be described poorly from training data and yet be found and named when an assistant searches. The reverse can happen too.
We saw how much this can matter in our published example report, from 23 October 2026. Asked 12 buyer questions in its category, Gemini without web search named the brand in 0 of 12 answers, while Perplexity, which searches, named it in 10 of 12. That's one sample, but it shows that "what the model knows" and "what the assistant says when it searches" can be very different numbers.
Which sources shape the answers
The grader's dimensions refer to the quality and authority of sources, but a one-off grade doesn't list which sites the assistants cited for your buyers' questions. That list is often the most actionable part of an AI visibility check, because it tells you where to get listed or corrected. In our example report, the most cited site was a HubSpot marketplace listing, in 21 of 48 answers, while the brand's own site was cited once.
Claude, and other assistants
The grader checks three platforms. It doesn't include Claude or Microsoft Copilot. If your buyers use those, you'll need to check them separately. In our example report, Claude named the brand in none of its 12 answers while ChatGPT named it in 9, so one assistant's result doesn't stand in for another's.
Change over time
Any single check is a snapshot. AI answers vary from one day to the next, so the useful signal is the trend across repeated checks with the same questions. HubSpot's own page makes this point when it positions HubSpot AEO for ongoing tracking.
How to use your grade well
- Run it for yourself and two or three competitors. The relative scores tell you more than the absolute number.
- Read the written interpretation for errors. If an assistant describes your products, markets or size wrongly, that's a fact problem you can fix on your site and on the sources that mention you.
- Then test the buying questions. Ask assistants the category questions your buyers ask, and note who is named and which sources are cited. Our free AI visibility checker does this: it asks four assistants, including Claude, 12 realistic buyer questions in your category, says which were asked with web search, and reports mentions, competitors and cited sources as a sample of what they said.
- Decide on tracking only if there's a gap worth closing. HubSpot AEO and the other tools in our AI visibility tools comparison all track over time; pick by the assistants and prompts you need.
The short version
HubSpot's AEO Grader is a good, free way to see how three AI models describe your brand from what they learned in training, scored on sentiment, presence, recognition, share of voice and positioning. It's not designed to show whether assistants recommend you to buyers who don't name you, what they find when they search, or which sources they rely on. Use it as one view, alongside a check of real buying questions.
For what to do with the results, read what is AEO and how to get cited by ChatGPT. If you're a HubSpot customer or partner and want help, our AI visibility audit and retainer cover the full check and the fixes, and you can book a free call to talk it through.