Expert Take: The SEO industry learned to game algorithms — in the era of hyperpersonalisation, we need to reinvent that game.
Last updated: April 2026
What do AI search engines actually do when someone asks a question?
Say a potential customer asks ChatGPT: “What’s the best contract management solution for mid-sized companies?”
You might expect AI search engines to search for that exact phrase. They don’t.
What actually happens: AI search engines translate that one question into multiple simultaneous queries. Something like “contract management software comparison”, “digital signatures mid-sized company”, “contract management cost per user”, “which companies use contract management software”. Those queries go to a search engine — in ChatGPT’s case, that’s Bing — and the answer is assembled from the combined results.
This is called query fan-out, and it changes what a brand needs to do to appear in that answer.
For B2B marketers, this is especially relevant. Gartner research shows that B2B buyers have completed an average of 57% of their purchase process before contacting a vendor. A large part of that research is invisible to you. AI search engines are now a fixed part of that early phase. By the time someone calls your sales team, that same person has already asked questions on ChatGPT or Perplexity — and already formed a shortlist. If you’re not in it, you were never an option.
Why optimizing for one keyword no longer works
The classic SEO reflex is: find the most-searched keyword in your category and build a page around it. For AI citations, that approach is insufficient.
AI search engines don’t reward pages that perfectly answer one question. They reward pages that answer multiple sub-questions simultaneously. A page that covers only “contract management software comparison” loses to a page that also addresses cost, implementation time, integrations, and ease of use — because that second page is relevant to more of the sub-queries.
That sounds like a technical puzzle, but the practical conclusion is simple: write more extensively about your subject than you’re used to. Not broader in the sense of vague, but broader in the sense of complete. A buying guide outperforms a product page. A comparison article outperforms a feature overview.
How to concretely get your brand into those answers
There are two things you can do without any technical knowledge.
Make sure your brand is in the text, not just in the metadata. AI search engines cite specific sentences from your content. Your brand name and specific features are pulled directly from the text. If you offer software with “automated reminders at contract expiry” or “integration with Salesforce”, those words need to literally appear in your content — not just as a bullet in a feature list, but as part of an explanation.
Write as if someone is asking follow-up questions. AI search engines look for sentences that directly answer a specific sub-question. If you include the question “How long does implementation typically take?” as a subheading in an article, followed by a concrete answer, you increase the chance that specific passage gets picked up. You can do this for every question a potential customer has during your buying process.
Write for everyone who has a say, not just the end user. In B2B, rarely does one person make a purchase decision. The CFO wants to know what it costs and what the ROI is. The IT manager wants to know how it integrates with existing systems. The end user wants to know what daily use looks like. All those people ask different questions to AI search engines — and all those questions deserve an answer in your content. If you only write for the person who ultimately uses your product, you’re missing half the DMU.
How do you find out what AI search engines cite you — or your competitor — for?
You don’t have to guess.
If you have a ChatGPT subscription, you can download your conversation history as a JSON file. That file contains the exact search queries that were executed. That gives you direct insight into which sub-questions AI search engines use for your category.
A more accessible method: look at which articles are currently being cited when someone asks questions in your product category. Paste the URL of such an article into an SEO tool like Ahrefs or Semrush. You’ll see which search queries that article ranks for — and those are exactly the sub-questions AI search engines consider relevant. That’s your content agenda.
B2B is an advantage, not a disadvantage
Many B2B marketers think their niche is too specific to seriously pursue AI visibility. That’s the wrong reasoning.
Precisely because B2B categories are narrow, there’s little good content available for AI search engines to cite. In a category like “project management software for construction” or “compliance solutions for financial institutions”, most pages are product-focused and shallow. A buying guide that genuinely answers your target audience’s buying questions stands out immediately. The bar for getting cited in B2B is considerably lower than in broad consumer categories — but most B2B marketers don’t try.
FAQ
Do I need to rewrite my entire website?
No. Start with the pages closest to the purchase decision. Comparison pages, buying guides, and FAQ sections deliver the most return.
Does this work for lesser-known brands?
Yes, as long as the content answers the sub-questions AI search engines are looking for. Brand awareness plays a smaller role than with traditional search engines.
How quickly do you see results?
Honestly, it’s hard to say. AI search engines don’t index on a fixed cycle. Expect weeks to months, not days.
Do I need to buy special software?
Not necessarily. You can research a lot manually with tools you probably already have. Paid tools are only useful if you want to do this at scale.
What if my competitor is already cited a lot?
Then you know what works in your category. Analyze what content they have, what questions they answer, and write something broader and more in-depth. That’s the most direct route.
Do I need separate content for each role in the DMU?
Not necessarily separate pages, but separate sections. A buying guide that covers both the business axis (cost, ROI, implementation time) and the technical axis (integrations, security, management) addresses multiple roles at once. That’s precisely the type of content AI search engines cite because it answers multiple sub-questions simultaneously.
Read also: Everything a B2B Marketer Should Know About GEO in 2026 · What is Entity Optimization? · What is the Knowledge Graph?