I ran into this problem around October 2025. We'd been building traffic steadily — decent Search Console numbers, a few solid backlinks, a content calendar we actually stuck to. Then a customer told me they'd found us because "ChatGPT recommended it."
I asked how often that was happening. He said that's how he evaluates every new tool now — he asks Claude or Perplexity first, then decides whether to click through.
I went straight to ChatGPT and typed our primary buying query. The one our homepage is built around. The one we'd published six articles targeting.
We weren't in the answer. Three competitors were.
That was the moment I realized we'd been optimizing for the wrong search engine.
#The shift that's already happened
There's a common version of this argument that frames AI search as a future threat. A "trend to watch." Something to prepare for in 2027.
That framing is wrong. This isn't coming — it's here, and buyers in B2B software are ahead of the curve on it. The profile of a buyer who uses ChatGPT or Perplexity as a starting point for tool evaluation is almost exactly the profile of a buyer with high intent and budget. They're not casually browsing. They know they have a problem and they're actively assembling a solution.
Perplexity crossed 15 million daily active users in early 2026. ChatGPT processes over 100 million queries per day. The category of query that's grown fastest — according to Perplexity's own published data — is research and product evaluation. "What's the best X for Y" and "compare these tools" queries.
Those are your buyers. And the answer they get is determined by which brands the AI treats as credible authorities on your category.
#Why being "on Google" doesn't mean being "in AI search"
This is the part that trips most founders up. You might have a page one ranking for your target keywords. You might even have a Featured Snippet. None of that guarantees you appear in AI-generated answers — and in some ways, a clean Google ranking can mask the AI visibility gap entirely.
AI models don't scrape your ranking. They cite what their training data and retrieval systems treat as authoritative. That means:
Volume of mentions across sources matters more than rank. An AI model that's seen your brand mentioned on G2, in a Hacker News thread, in two industry newsletters, and in a Capterra comparison — that's more likely to cite you than a brand with one top-ranking page and nothing else.
Structured content that directly answers questions gets weighted more heavily. FAQ sections, comparison tables, clear definitional paragraphs — these map to how AI models compose answers. Dense prose that's optimized for humans reading linearly doesn't help the model extract a clean citation.
Recency matters for retrieval-augmented models. Perplexity and ChatGPT with Browse are pulling from live indexes. A post published last month beats a technically superior post from two years ago.
#The three-layer fix
I'm not going to tell you this is easy to get right quickly. It's not. But the components aren't mysterious — they're just different from what most SaaS marketing teams are spending their time on.
Layer 1: Content that actually answers questions
Not "content that ranks for keywords." Content structured to answer the specific question someone might ask an AI about your category.
What's the difference? A page optimized for "project management software for remote teams" might look like a landing page — headline, features, CTA. That's not what an AI cites. What gets cited is a page that explicitly answers "what makes project management software work for remote teams?" with clear, structured answers — ideally in FAQ format, with specific criteria, with named examples.
The content strategy shift: write for questions, not keywords. Your FAQ pages are now your most important SEO assets, not your lowest-priority afterthoughts.
Layer 2: Structured data that AI retrieval systems can parse
Every SaaS product page should have SoftwareApplication schema with a clear category, description, and pricing. Every FAQ section should be accompanied by FAQPage schema. Your Organization schema should include sameAs links to every platform where your brand has a presence — G2, Capterra, LinkedIn, Product Hunt, Crunchbase, GitHub.
That last point matters more than most founders realize. The sameAs property tells every machine-reading system that all these references point to the same entity. It's how AI models build confidence that your brand is real, established, and worth citing.
Layer 3: Third-party corroboration
Your own website can't be the only place that talks about you. AI models weight corroboration — the same brand mentioned across independent sources is more credible than one that appears only on its own domain.
This means: genuine reviews on G2 and Capterra (20+ is the threshold where it starts making a difference), a real Product Hunt listing with an actual description, at least one mention in an industry publication that isn't a press release you wrote. Hacker News "Show HN" posts, relevant subreddit threads where you've added real value, developer community mentions if your product has a technical component.
None of this is new advice in isolation. But most founders I know are treating it as nice-to-have brand work rather than a core component of how buyers discover them.
#The measurement problem
Here's what I didn't have when I found out we were invisible: any way to track it systematically.
I was manually prompting ChatGPT and Perplexity with our target queries. Which works once. But you can't do it across five platforms, thirty queries, and twelve competitor names, every week, without burning hours you don't have.
That's the gap that dedicated AI citation tracking tools fill. Presence AI is the one I've spent the most time with — it monitors your citation rate across ChatGPT, Claude, Perplexity, Gemini, and others against a defined set of queries, benchmarks you against up to twenty competitors, and surfaces the specific queries where you're losing ground. The kind of thing that would take a half-day per week to do manually, automated into a dashboard you check like you check Google Search Console.
The specific data point that changed how I thought about this: seeing exactly which competitor was getting cited instead of us, for which queries, on which platforms. That's actionable in a way that "we need more AI visibility" never is.
#My actual take on urgency
I'm not going to tell you to drop everything and pivot your marketing strategy today. But here's the honest version of the risk calculus.
The brands building AI visibility now are embedding themselves in the training data and citation patterns that will become harder to displace over time. Not impossible — models retrain, retrieval indexes update — but there's a first-mover dynamic that resembles early SEO more than people want to admit.
The founders who built domain authority in 2010–2015 are still benefiting from it. The ones who started in 2018 had to work three times as hard for the same positions.
I don't know if AI search will follow the same curve. But given that I can see my customers arriving pre-sold because an AI model recommended us — I'm not treating it as a future problem.
The audit takes two hours. Start there.
For more on how AI is changing the discovery layer for SaaS products, see the Business & Marketing section.