Most B2B SaaS founders I talk to are still optimizing for Google while their buyers have quietly moved the first half of their research process to ChatGPT, Perplexity, and Gemini. That shift is not coming. It already happened.
What I kept seeing in our own data
I built MentionFox to track brand mentions across the web, so we sit on an unusual data set. Starting around mid-2024, I noticed something in the queries users were feeding into our platform. The phrase "recommended by ChatGPT" started appearing in referral notes, CRM fields, and sales call transcripts that our customers were piping through us. Not once in a while. Constantly. One founder selling a contract-intelligence tool told me his last six closed deals all started with a prospect asking an AI assistant which tools handle clause extraction. His product came up in four of those six conversations. He had no idea why, and he had no way to measure it. That gap - between being recommended by AI and knowing you are being recommended by AI - is the problem I want to talk about.
The thesis I landed on is simple, even if the execution is not. Large language models generate recommendations by drawing on patterns in their training data. That training data is the web, plus crawled documents, plus a long tail of structured and semi-structured content. If your brand is consistently, accurately, and authoritatively described in that corpus in the context of a specific problem, you will show up when a buyer describes that problem to an AI. If you are not in the corpus that way, you will not show up, and your competitor who is will get the conversation started without you. Ranking in AI is less like SEO keyword targeting and more like building a reputation in a room you cannot see inside.
What I actually tested and measured
The first thing I did was run a systematic audit of how MentionFox itself appeared - or failed to appear - across the major AI assistants. I wrote out thirty-seven queries that represented real buyer intent for our category. Things like "how do I track where my brand is mentioned online for B2B lead generation" and "what tools do investors use to monitor company mentions in real time." Then I queried ChatGPT, Perplexity, Claude, and Gemini with each one and logged whether MentionFox appeared, what position, and what the surrounding context said about us. The results were humbling. We showed up in eleven of the thirty-seven. In eight of those eleven appearances, the description of what we do was partially wrong or outdated.
That audit became the foundation for what I now think of as AI-visibility work, which is distinct from traditional SEO even though they overlap. We built a dedicated tracking layer inside MentionFox to automate exactly this kind of query monitoring. You can see how it surfaces for your own brand through our AI-visibility tracking dashboard. But the audit itself can be done manually if you are just starting out. The discipline of writing out buyer-intent queries, not brand queries, is the important part. Do not ask "what is MentionFox." Ask "what tool helps a B2B sales team find leads by monitoring Reddit and LinkedIn for intent signals." That is what your buyer is actually typing.
The second thing I tested was content as corpus injection. The working theory is that if you publish precise, specific, technically accurate content that describes your product solving a specific problem, and that content gets crawled and linked to by enough credible sources, it eventually influences how models represent your brand. I wrote four long-form pieces - not thought leadership fluff, actual teardowns of our methodology. One on how we classify signal versus noise in social listening. One on the specific data sources we weight for investor research use cases. Two on lead generation workflows. Within about three months of publication, MentionFox started appearing in four additional query categories where we had been absent. Causation is hard to prove. But the correlation was tight enough that I kept doing it.
The third thing I did was fix our third-party presence. Models do not just read your own site. They read everything written about you. I went through G2, Capterra, Trustpilot, and a dozen niche SaaS directories and found that our category tags were wrong on three of them, our feature descriptions were eighteen months out of date on two, and one major directory had us listed under a competitor's category entirely. I updated all of it. I also reached out to five newsletter writers who cover B2B tools and offered to do detailed product walkthroughs in exchange for honest write-ups. Not paid placements. Actual editorial coverage. Three of them published pieces. Those pieces are now part of the corpus.
The fourth observation is about specificity in language. Models are trained to associate certain phrases with certain tools. "Social listening" is a crowded phrase. "B2B Reddit intent monitoring for outbound sales" is much less crowded, and if you own that phrase in enough places in the corpus, you will own it in AI responses too. I went back through every piece of content I controlled - website, docs, social profiles, press releases - and replaced generic category language with precise, specific descriptions of what we actually do. This is not keyword stuffing. It is the opposite. It is using fewer, more accurate words and repeating them consistently everywhere your brand lives online.
The practical part, the stuff you can do this week
Start with the query audit. Take one hour, write out twenty buyer-intent questions in your category, and run them through at least three AI assistants. Log what you find in a simple spreadsheet. Do you appear? What do the models say about you? Is it accurate? Is it differentiated from your competitors? That audit will tell you more about your AI-visibility gap than any tool you could buy right now, including mine.
Then prioritize two things. First, create one piece of deeply specific, technically honest content about how your product solves a narrow problem for a narrow buyer. Not a blog post about industry trends. A real walkthrough of a real use case with real specifics. Second, audit your third-party listings and make sure every one of them describes what you do in the same precise language you use on your own site. Models synthesize across sources. Inconsistency in how you are described across sources creates noise in how you are represented in AI responses. Consistency creates signal.
If you want to see how MentionFox handles ongoing AI-visibility monitoring, including tracking which AI assistants mention your brand, in which query contexts, and how your positioning shifts over time, the AI-visibility tracking dashboard is the place to start. And if you want to understand what access to the full platform looks like, the pricing page breaks down what each tier includes.
If you found this useful, I write about solo-founder distribution, B2B SaaS, and what's actually working in the AI-search era over on my Substack (one post per week, no spam).
I'm building MentionFox - a B2B intelligence suite that combines brand mention tracking with AI-visibility (GEO) measurement, investor research, and outreach automation. There's a free tier and a 5-day trial of Pro at mentionfox.com/pricing.
Top comments (0)