We are a small B2B company, and this is an honest account of eight weeks spent using HubSpot's beta AEO tool to get cited by ChatGPT, Gemini and Perplexity for one specific service. Our visibility score moved from 8.62% to 26.63%, and most of that progress came from decision-stage prompts, not the awareness-stage ones we expected to matter. We are still adjusting as we go, and we suspect most companies trying this will be too.
We'll be honest. When HubSpot first announced the AEO tool, our reaction was somewhere between "finally" and "here’s another tool to figure out." We already know that AI tools are changing how people search, and as a small B2B company, we were wondering how visible we were and if we were showing up for the things that mattered.
So we decided to poke around. Everything in this article comes from our own experience of using HubSpot's new AEO tool over eight weeks, starting in mid June 2026. Nothing here is theoretical. We ran the experiment on ourselves before we would think about recommending it to anyone else.
One thing worth clarifying before anything else, because people sometimes use the terms interchangeably: a citation and a brand mention are not the same thing. An AI tool can quote or draw information from your website without ever naming your company in its answer. Someone might ask, "How do I map data correctly when migrating from one CRM to another?" and get an answer built partly from our content, but without our name being cited. This isn't something we discovered, it's simply how these systems work, and we saw it play out clearly, right at the start of our experiment, when we noticed one of our own old blog posts being drawn on for an answer that never mentioned us at all.
People have started asking AI tools the questions they used to type into Google, and increasingly, even inside Google, they are using AI mode more. By the time someone actually visits a company website, they have often already formed an opinion about who the credible options are. If a buyer asks an AI tool something like "who are the best Salesforce to HubSpot migration partners in India?" we would obviously like our brand to come up in the answer.
For a services business like ours, that matters commercially. We've been doing SEO work for the better part of ten years now, long enough to have gone through a few of these "the rules are changing" moments before, and AEO is the next one.
As a HubSpot Platinum Partner, we had access to the AEO tool the moment it launched in April 2026, and so did our Marketing Hub Pro clients, who were already asking us about it. We wanted to be able to honestly say that we had used it ourselves, seen what worked and what didn't, and formed our own view rather than repeating a vendor's pitch back to them. We have been there and done that, and that counts for something when a client asks whether this is worth their time.
HubSpot built the AEO tool to help businesses track and improve how they show up in AI-driven discovery, after HubSpot's own customers reported a 27% year-over-year drop in traditional organic traffic. The tool tracks brand sentiment, competitor share of voice and source citations across AI answer engines, including ChatGPT, Gemini and Perplexity.
What makes it different from a generic tracking tool is that it uses a business's own CRM data to generate search prompts that mimic how real buyers actually research, rather than making you guess at the right questions yourself. Once it has identified where your visibility is weak, it gives you prioritised recommendations, campaigns to run, content to publish, and pages to edit, all of which you can act on directly inside HubSpot.
So, here's what happened when we tested it on ourselves.
Our first instinct was to track three services at once. We talked ourselves out of it fairly quickly, because comparing three moving targets at the same time sounded like a good way to learn nothing clearly about any of them. We settled on Salesforce to HubSpot migration: a service we are genuinely good at, one where we occasionally showed up on the first page of Google but never consistently, and one we wanted more enquiries for.
The tool's auto-generated prompts didn't quite fit our services as they stand today, so we wrote our own. We drafted 15 prompts, whittled them down to 12 with input from our HubSpot and SEO teams, and loaded them into the tool.
We began in early June, with our overall brand visibility sitting at 8.62%, shown in red on the dashboard. HubSpot's own guidance, and the tool itself, warned us that it would take time to produce anything meaningful, so we settled in to watch and wait rather than panic. After two weeks, our visibility was at 15.41%
Two things we noticed immediately. Gemini cited us more than the other tools by a significant margin, while ChatGPT did not acknowledge our existence for any of the prompts. That was odd, because we already knew we ranked on page one or two of Google for this exact service. So, where was ChatGPT getting its information from?
We went looking at the sources each tool was actually drawing from: a mix of peers, competitors, review sites, user-generated content, and our own pages. What we found was a narrative about our company that was inconsistent across every one of those sources. Somewhere in that research, we came across a company blog page on Medium that we hadn't touched in years. It was being cited as if it were current. Nobody on the team even remembered it existed. This was also the moment we saw, first-hand, a source being used without our brand name attached to it.
We also found ourselves listed on review sites we had never actively managed or even registered for. So we did the necessary groundwork of rewriting our narrative consistently across these review sites and cleaning up the story wherever we could influence it. We also updated our HubSpot ecosystem partner profile. All this activity took us to the end of June.
The results came quickly once the narrative work was done. Our visibility score climbed to 18.40% (still red, but moving), and kept climbing week on week to 26.63%, taking us from red into orange.
As can be seen from the screenshots below, ChatGPT, which had barely mentioned us at the start, began citing our brand more often, and so did Perplexity.
Prompt One. Run on 7th July.
Prompt One. Run on 17th August.
Prompt Two. Run on 7th July.
Prompt Two. Run on 17th August.
Two blog posts written from the tool's topic recommendations helped us rank in Google's AI overviews for a few different prompts.
And on the commercial side, enquiries for our Salesforce to HubSpot service increased noticeably, though we are still tweaking things as we go rather than declaring victory.
The tool groups prompts into awareness, consideration and decision stages, and recommends spreading your prompts across all three. In our case, none of our awareness-stage prompts produced a citation, despite following the recommendations attached to them. What this taught us, for our own use going forward, is to keep awareness-stage prompts to a minimum, put a moderate number into the consideration-stage, and weight the majority towards decision-stage prompts. That balance worked for us. It may look different for someone else.
If you are in B2B, you might want to keep in mind that awareness-stage prompts are crowded, and being cited there is much harder than most people expect. When someone asks a specific question and gets a useful answer, they move on. They rarely check who was quoted, unless the question is high-stakes, sensitive, or one where authority genuinely matters to them. Our name only reliably appeared at the decision stage, where someone has already narrowed things down and wants a specific answer.
We also had to question our own prompt choices. We picked prompts we assumed buyers were asking, validated by our own team, but that isn't quite the same as knowing what people are actually typing into these tools. That's a different research problem, and probably needs a different tool to solve properly. We will let you know when we crack that.
There were gaps in what the tool covers, too:
At the end of the day, it is a tool, and it does not replace your judgement. The tool won't tell you to check what your competitors are doing, so you can try that too. You have to do your research, connect the dots, pick the prompts that work for you, and decide which recommendations are worth acting on.
For us, a lot of that comes down to instincts we've built over years of SEO work, knowing which signals actually matter and which are noise. You still have to do that thinking yourself, and honestly, tool or no tool, all the above is exactly what we would have told a client anyway.
AI visibility is not an exact science, and 8 weeks of data isn't enough to claim we've solved it. Our score still fluctuates week to week, and we've come to expect it rather than worry about it. You are not going to show up every day in every AI tool, and that's normal too.
We are continuing to run this experiment on ourselves, adjusting the prompts, refining what we publish, adding more services, and paying closer attention to the sources feeding these answers, and we'll keep sharing what we find as we go.
If you've been running your own experiments with AI visibility, or you're a HubSpot Partner comparing notes, we'd genuinely like to hear what you're seeing. And if you'd like to explore HubSpot's AEO tool for your own business, whether as a standalone product or as part of a Marketing Hub subscription, we're happy to walk you through it. Feel free to connect with us.