Before a prospect ever reaches your site, an AI assistant has already described your company, compared you to rivals, and decided whether you make the shortlist. Here's how to make sure it gets you right.
In one line: 94% of B2B buyers now use generative AI somewhere in their buying journey. The company that gets recommended is the one AI can describe clearly and verify against trusted sources, which is not always the best one on the market. Generative engine optimization (GEO) is how you become that company.
Try this before reading on. Open ChatGPT and ask it about your company. Read what comes back. That paragraph is now doing the job your homepage used to do, and you didn't write a word of it.
Most founders and marketing leaders we speak to have never run that check. A Forbes Agency Council piece by Nital Shah put a hard number on something we see in almost every AI visibility audit: by the time a buyer books a demo, a machine has already introduced you, summarized what you do, and decided whether you're worth a look.
A machine now makes your first impression
The number comes from 6sense's 2025 B2B Buyer Experience Report: 94% of B2B buyers use generative AI at some point in their buying journey. The AI doesn't sign the contract, a human still does that. It decides who gets past the velvet rope and into the room where the contract gets discussed.
So a machine is making your first impression, and almost nobody has checked what it says. That is the gap, and closing it has a name: generative engine optimization.
Discovery has moved inside the AI
For two decades the B2B buyer journey ran on a familiar loop: search, compare, shortlist, reach out. Much of that discovery now happens inside AI assistants instead. The same 6sense report found that buyers build their shortlist early and go on to buy from that initial list roughly 95% of the time. If AI leaves you off the list, you may never learn you were in the running.
This is already measurable, not theoretical. Vercel reported that ChatGPT went from referring under 1% of new sign-ups to around 10% in about six months. AI search optimization stopped being a 2027 problem some time ago.
AI recommends only what it can see
An AI assistant doesn't invent its recommendations. It assembles them from articles, reviews, directories, comparison pages, analyst reports, and community discussions, the sources it has learned to trust. The pattern we hit again and again during AI visibility audits is blunt: strong companies are simply absent from those sources, while rivals with heavier third-party coverage show up in answer after answer.
The company that gets recommended is the one AI can see clearly and cross-check, and that is not always the best product on the market. You cannot be recommended if the model never learned enough about you to feel safe naming you.
Confident, and often wrong
Here is where it turns risky. An LLM sounds equally certain whether it is current or eighteen months stale. It blends its training data with trusted web sources and, even when it browses live, tends to weight authoritative third-party sites above your own marketing pages.
We watched this play out with a B2B SaaS company that had changed its pricing model. More than a year later, the leading assistants were still quoting the old numbers. The website was correct. The AI was not, because directories and partner sites still carried the previous figures and outranked the company's own pages in the model's eyes.
The damage went past a wrong answer. It became a trust problem. Prospects arrived certain the AI's pricing was right, so when the sales team quoted the real figure, buyers assumed the rep was upselling them. Every deal now opened with the team defending a version of the company that no longer existed.
The GEO playbook: 4 moves to become legible to machines
None of this needs a large AI budget. The brands pulling ahead are the ones treating their own information as an asset worth maintaining. Four moves do most of the work.
1. Audit your AI presence. Ask ChatGPT, Gemini, and Perplexity about your company, your products, and your competitors. Read every answer as if a prospect wrote it, then log what is wrong, missing, or out of date. Run the same AI visibility audit every month, because the answers keep moving.
2. Tell one consistent story. Your website, LinkedIn page, G2 profile, partner listings, press coverage, and directory entries should all describe the same business. AI rewards consistency long before it rewards creativity, and mixed messages turn into mixed answers.
3. Earn third-party validation. Reviews, analyst mentions, industry publications, trusted directories, and earned media shape how machines describe you. Those assets have quietly stopped being only PR. They are training data now.
4. Measure what actually matters. Traffic is a fraction of the picture. Track whether AI recommends you, how often you get cited, whether your positioning comes through accurately, and how AI-referred visitors behave once they land. Adobe found that shoppers arriving via AI over the 2025 holiday season converted 31% better than other channels. Fewer clicks does not mean fewer customers.
The Bottom Line
Two decades of digital marketing were spent helping humans find information. The next decade is about helping machines understand it. The brands that win this shift won't be the ones with the biggest AI budget or the cleverest prompts. They'll be the ones whose business is easiest for AI to understand, verify, and recommend. That is discipline more than creativity, which is good news, because discipline is a decision any team can make this quarter. Most companies haven't even asked AI what it thinks of them yet. Start there.
Key figures referenced in this piece
- 94% of B2B buyers use generative AI at some point in their buying journey (6sense, 2025)
- Buyers purchase from their initial shortlist ~95% of the time (6sense, 2025)
- ChatGPT referrals to new sign-ups grew from under 1% to ~10% in about six months (Vercel)
- AI-sourced shoppers converted 31% better than other channels over the 2025 holiday season (Adobe)
Blog post inspired by the Forbes article "Your Next Customer Is A Machine: Marketing When 94% Of B2B Buyers Ask AI First" by Nital Shah, Forbes Agency Council.
Find out how AI sees you, before your next lost customer does
Krein's GEO Audit Tool analyzes content, structure, and digital signals to measure your brand's presence within Generative Engine Optimization ecosystems. Not assumptions, butconcrete data on what AI systems say about you, where the visibility gaps are, and which priorities to address first. Try the GEO Audit Tool and see how your B2B company is currently positioned, cited, and described by the leading LLMs.
FAQ
What is generative engine optimization (GEO)?
GEO is the practice of making your business easy for AI assistants like ChatGPT, Gemini, and Perplexity to understand, verify, and recommend. It extends SEO from ranking on search results to being accurately cited inside AI-generated answers, mainly by earning consistent, trusted third-party coverage.
Why do B2B buyers ask AI before visiting a website?
Because it is faster. Instead of searching, comparing, and shortlisting manually, buyers now ask an AI assistant to summarize a market and suggest vendors. With 94% of B2B buyers using generative AI in their journey, the AI's answer often forms the shortlist before your own site is ever opened.
How do I check what AI says about my company?
Run an AI visibility audit: ask ChatGPT, Gemini, and Perplexity about your company, products, and competitors, and read the answers as a prospect would. Note anything inaccurate, missing, or outdated, and repeat monthly, because the answers change over time.
Why does AI recommend competitors instead of us?
AI recommends what it can see in trusted third-party sources, reviews, directories, analyst reports, and publications. If competitors have stronger external coverage and you don't, the model has more evidence to name them, regardless of which product is actually better.