For years the plan for being found was simple to describe: rank in Google for the questions your buyers search, and turn the clicks into leads. Now a buyer types the same question and gets a written answer at the top of the page, or asks ChatGPT and never visits a search engine at all. Founders want to know how a B2B company still gets found, and how it ends up in the answer.
The worry is practical. A founder on Hacker News calls SEO and GEO the best lead channels while describing the effort as a real pain (Hacker News). Another thread asks simply whether SEO traffic is up or down (Hacker News). And the business stakes are visible: one widely shared thread linked layoffs at Tailwind to falling documentation traffic as developers turned to LLMs (Hacker News).
What works is a fuller presence: a site that search engines and assistants can read and trust, mentions of your company in places you do not control, and measurement that tells you what is changing.
What the numbers say about clicks and shortlists
Clicks fall when a summary answers the question. Pew Research Center tracked 68,879 searches by 900 US adults and found that when an AI summary appeared, users clicked a traditional result in 8% of visits, against 15% when there was no summary (Pew Research Center). Users clicked a link inside the summary itself in only 1% of visits. Summaries appeared on 18% of searches overall, rising to 60% of searches phrased as questions, which is exactly how B2B buyers tend to search.
Ahrefs estimates that AI Overviews were associated with a 58% lower click-through rate for the first-ranked result than forecast, across 300,000 keywords (Ahrefs). It is a correlational, forecast-based estimate, but it points the same way as Pew.
At the same time, buyers are moving their research into assistants. G2's 2026 survey of 1,076 software buyers found that 51% start research with an AI chatbot more often than with Google, and that 69% chose a different vendor than planned because a chatbot recommended it (G2). Forrester reports that 94% of business buyers use AI during their buying process (Forrester). G2 has a commercial interest in buyer research, but the Forrester figure confirms the broad shift.
Classic results now send fewer clicks while AI answers shape more shortlists, and your plan has to serve both.
Start with what Google says it needs
Google's own documentation is plain on this point. There are no additional requirements or special optimisations needed to appear in AI Overviews or AI Mode; a page must be indexed and eligible to show in Search (Google Search Central).
That puts more weight on the technical basics. Pages must be crawlable, load quickly, render their main content without heavy scripts, and sit in a clear structure with sensible internal links. A surprising number of B2B sites hide product detail behind tabs, gated PDFs or images of text that neither search engines nor assistants can read well.
Founders on Hacker News experiment with structured data and with licence notes addressed to AI models, without clear evidence yet that these help (Hacker News, Hacker News). Structured data is worth doing properly for products, organisations and articles, because it helps search engines understand the page. Just do not expect it to replace clear content.
Make your pages easy to quote
Assistants and summaries draw on pages that answer a specific question clearly. Write for the questions your buyers ask in the middle of their research: what the product does for a given role, how it compares with the usual alternatives, what it costs, what it integrates with, how long setup takes, and what security and compliance look like.
Answer each question near the top of its page in plain sentences, then add the detail. Use real numbers you can stand behind, named integrations and concrete examples. Vague category language gives a model nothing to repeat, while a specific statement such as who the product is for and what it replaces is easy to cite.
Then build presence outside your own site. G2 found that AI chatbots are now the top source influencing buyer shortlists, ahead of review sites, analysts and vendor websites (G2), and those chatbots learn about vendors from the wider web. Reviews on the platforms your buyers use, honest participation in the communities where they ask questions, partner pages, podcasts and trade press all give assistants independent evidence that your company exists and what it is good at. Founders on Indie Hackers share their own playbooks for getting mentioned by ChatGPT and Perplexity (Indie Hackers), and the common thread is consistent, specific presence across many sources.
Measure what you can, honestly
Measurement here is imperfect. Google says traffic from its AI features is included in overall Search Console performance reporting (Google Search Central), so you will not see it as a separate line. In your analytics, set up a channel group for referrals from assistants such as ChatGPT, Perplexity, Gemini and Copilot, and watch that traffic and its conversion rate monthly.
The visitors may be worth more than their numbers suggest. Semrush's own benchmark estimates that the average visitor from AI search converts at around 4.4 times the rate of an average organic visitor (Semrush). Treat that as a vendor estimate, and check it against your own data.
For visibility inside answers, run a fixed set of buyer prompts each month across the main assistants and record whether you are mentioned and how. Be careful with personalised results. One founder on Hacker News believed their product was recommended in an AI Overview, and a reply pointed out that results are personalised and should be checked in an incognito window (Hacker News).
What to do this week
- List the 20 questions your buyers ask mid-research and check whether your site answers each one in plain text near the top of a page.
- Confirm your key pages are indexed in Search Console and that their main content is readable without scripts, tabs or gated files.
- Add an analytics channel group for AI assistant referrals and note the current baseline.
- Run ten buyer prompts in ChatGPT, Perplexity and Gemini in a logged-out or private session, and record who gets mentioned.
- Pick two third-party places where your buyers gather, such as a review platform or community, and plan steady, useful participation.
Being found now means being understood by people and by the systems that answer for them. Most of what it takes is familiar: a clean, fast site, clear and specific answers, real evidence and a presence beyond your own domain. Stolos Digital builds B2B websites with this structure in mind, from technical foundations to the pages buyers and assistants read first.