Rethinking “being found” when AI answers before people click
Google’s AI now regularly thanks brands for their content, then hands the click to a competitor. For an expert-led small business, that is not a clever quirk of technology. It is lost revenue dressed up as innovation.
In one review of 100 B2B “best [category] software” searches, AI Overviews pulled self-promotional listicles into the answer again and again. Out of 80 prompts that triggered an overview, those pages were cited more than 300 times. Yet in roughly 69 percent of those cases, the brand that wrote the article was not recommended. Their content trained the answer engine, their rivals took the lead position.
There is a vivid example. For the query “best LMS for selling courses”, an LMS vendor’s own listicle supplied the raw material. The overview then recommended four other platforms named in the article and quietly skipped the author. Helpful, just not to the business that paid for the content.
For expert-founders, that exposes the real game of answer engine optimisation. It is no longer enough to shout “we are the best” on owned content. AI search behaves more like a sceptical buyer. It prefers brands that are cited, reviewed and discussed in many places, not just on their own blog.
At the same time, owned content still sits at the centre of the system. AI agents and search engines continue to crawl websites and blogs. Expert-led firms that publish clear, current answers to real questions give both humans and machines a reliable source of truth. When that content lives on a site the business controls, it also becomes the safest place to guide visitors from curiosity to conversation.
Three patterns keep showing up. First, self-ranked “best of” pages can quietly damage both traffic and trust, especially when every list magically puts the house brand first. Second, external validation matters more than ever. Brands that show up in independent articles, reviews, videos and community threads are far more likely to be recommended, not just cited. Third, neglected archives of old, vague posts tell AI models that the expertise is fuzzy or out of date.
Designing answer engine optimisation for an expert-led small business starts with a different brief. Each article should answer a specific, high-intent question in plain language, feature the founder’s real point of view, and link out to credible third parties who also talk about the issue. Each quarter, teams can prune or refresh content that no longer reflects how they work. Alongside that, they can pursue reviews, interviews and guest appearances that create independent signals pointing back to the same expertise.
The practical test is simple. For the key questions that should lead to the business, do answer engines see a consistent, human, question-led story across the website and the wider web, or just a stack of self-congratulation? Expert-led brands that get this right will not only train AI to recognise them. They will also feel far more like the obvious, trusted choice to the humans who still make the final decision.
This content was co-authored by Draiper co-founder Tim Brown in collaboration with Draiper ContentFlow, a human-in-the-loop, AI-powered content workflow assistant. The final result was produced from idea to finish in under 3 minutes.