HubSpot's AI Search Grader: what it measures and how to use it

A tool like this exists because the place people look for answers has moved. A growing share of category research now happens inside ChatGPT, Google's AI Overviews, Perplexity, and similar assistants, where the model writes a direct answer instead of returning a list of links. Classic SEO tools track where a blue link sits on a results page, so they have no way to tell you whether an assistant named you when it answered a buyer's question out loud. The AI Search Grader reads that newer layer, which is why it stands as its own tool.

A note on scope: the grader measures the AI-answer layer specifically, and it captures where you stand at the moment you run it rather than tracking you day to day. We cover AEO graders as a general category in a separate piece in this hub; here we're staying with HubSpot's tool in particular.

VIDEO TRAINING

Get the Growth Playbook.

Learn to plan, budget, and accelerate growth with our exclusive video series. You’ll discover:

• The 5 phases of profitable growth
• 12 core assets all high-growth companies have
• Difference between mediocre marketing and meteoric campaigns

What does the HubSpot AI Search Grader measure?

The AI Search Grader measures how visible and how accurately represented your brand is when AI assistants answer questions in your category. In practice that breaks into a few related readings: whether the models mention you at all, how favorably and accurately they describe you when they do, and which competitors the models name alongside or in place of you.

Each of those readings points to a different kind of work, which is why the detail underneath the headline result tends to be the part worth studying more closely than the summary number on its own. If the models never mention you, you have a coverage problem, and the work is publishing genuinely useful answers on the questions you're currently absent from. Where the models do mention you but describe you vaguely or with stale information, the issue is accuracy instead, so the work shifts to clearer, better-sourced pages that state the facts you want repeated. And when you're losing ground to a specific competitor, that's a positioning problem, which usually means studying what that competitor publishes on the winning questions and then covering the same ground with more useful detail.

Here is how the main checks map to why they matter and what you do about each one.

A single run captures all of this at one moment, which makes the AI Search Grader a diagnostic instrument for where things stand today. It reads that snapshot rather than following your visibility from one day to the next. Most of its value sits in what it tells you to go fix, so we spend more attention on the gap list and recommendations beneath the headline result than on the summary score itself.

How do you run the HubSpot AI Search Grader?

You run the AI Search Grader from HubSpot's website by entering your brand details and letting the tool query the models on your behalf. The tool is free and doesn't require a paid HubSpot subscription to try, so the practical first step is to search for "HubSpot AI Search Grader," open the tool, and give it the brand and category information it asks for.

Because HubSpot updates the interface and the underlying methodology as this space evolves, run it yourself and follow the on-screen prompts rather than memorizing a fixed set of fields. The general flow holds steady: you supply your brand and enough context for the tool to understand your category, it sends representative prompts to the AI models, and it returns a result with the visibility readings and a set of recommendations.

A couple of habits make the run more useful. Give the tool accurate, specific category context so the prompts it generates match the questions your real buyers ask, since a vague description tends to produce generic prompts that won't reflect how you actually get found. It's also worth saving the output from your first run, because that initial result is your baseline, and the readings only start to mean something once you have a second run to compare against.

How do you read your HubSpot AI Search Grader results?

Read the result as a relative baseline, since the visibility reading only starts to carry meaning once you set it next to your competitors and your own earlier runs. A standalone number tells you little on its own. Set that same number against two competitors who clearly outrank you in the answers, though, and it starts to show you how much ground you have to cover and on which questions you're losing it.

A few reads matter more than the summary. Start with your standing on the questions that sit closest to a purchase, because being absent from a high-intent question costs you more than being absent from a top-of-funnel one. Look next at how the models describe you, since a model that gets your facts wrong calls for a different fix from a model that has simply never encountered you, and knowing which of those you're dealing with tells you whether to correct an existing page or write a new one. The trend is the third read, and it only appears once you've run the tool more than once, which is why we use the first run as a baseline and read each run after it as evidence that the work is landing.

One caution keeps the result in proportion. AI answers vary between runs even for the identical prompt, because the models are probabilistic and their training data keeps shifting underneath them. A single run is best treated as one sample whose individual answers can occasionally mislead, which is why we read the pattern across the whole set of prompts and avoid over-reacting to any one answer. When the same gap appears across several related questions, that's a signal worth acting on.

What can't the HubSpot AI Search Grader tell you?

The AI Search Grader can't tell you why a model describes you the way it does, and it can't fix anything on its own. It reports the symptom, meaning you're absent from a question or described poorly, without diagnosing the cause, which usually lives in your content, your structured data, or your simple absence from the topic. Interpreting the result and doing the work remains your job.

Two limits are worth naming so the result doesn't get over-read. The tool samples a moment in time, so a result can wobble between runs without anything real having changed on your site, and that variance is a property of the models rather than a flaw in the tool. The category is also young, so HubSpot and every other vendor in this space refine their methodology as the models change, which means the same tool can read differently from one quarter to the next for reasons that have little to do with your own pages.

None of that makes the tool less useful. It finds the gaps quickly and for free, and the work of closing them is content and schema work that sits downstream of the result. We let the AI Search Grader set our priority order, then spend the real hours on the pages it points us toward. When we want a broader read on the same site, we pair the grader with our AI website teardown, which looks at the page-level structure underneath the visibility number.

How should you act on your HubSpot AI Search Grader results?

Turn the result into a ranked work list and start with the high-intent questions where you're absent or poorly described. The grader hands you a gap list and a set of recommendations, and the move is to sort that list by how close each question sits to a buying decision, then work top-down so your effort lands first on the questions most likely to produce a customer.

The sequence we follow with clients holds up across accounts. We begin by writing answer-first content for the questions where the models don't mention us at all, because a model can't quote you on a question you've never genuinely answered. From there we tighten the pages where the model does mention us but gets the description vague or dated, so the engine has accurate, current facts to repeat. Alongside the content we make sure the structured data is clean so engines can parse and attribute what we've published, and we re-run the grader on a regular cadence to confirm the visibility is actually climbing rather than assuming it is.

If you'd rather hand the whole loop to a team that already owns the stack, our AEO Authority System covers the auditing, schema, and content work end to end for HubSpot sites. For teams running it in-house, the principle is the same whichever tool you start with: let the result set the priority order, then put your real hours into the content and structured data that close the gaps it surfaced.

Schema markup recommendations

Two schema types do most of the work for a tool walk-through like this one. Mark up the question-based H2 sections with FAQPage schema so engines can pull each question-and-answer pair directly into an AI response, since every heading here is written as a real query someone would type into an assistant. Wrap the whole piece in Article schema with a clear author, datePublished, and dateModified, because authorship and recency are signals AI systems weigh when deciding which source to trust, and a topic tied to a fast-moving tool dates quickly enough that an accurate dateModified is worth maintaining.

Keep the comparison table as clean HTML <table> markup rather than an image, so engines can read the "what the grader checks" rows as structured data. If you publish schema across a HubSpot site and want it generated and maintained alongside your pages rather than hand-placed on each one, that's the job our machine-readable structured data module was built for. Whichever route you take, validate every block in Google's Rich Results Test before you publish, and confirm the visible content matches the markup so the structured data actually earns the citation.