Google Ranked Pages. AI Will Rank Evidence.
Dan Lievens on why technical SEO is only part of AI search visibility, and how specific, honest customer experiences can strengthen your evidence.
By Dan Lievens, co-owner of Share One. Originally published on LinkedIn on September 18, 2026.
Short answer: Technical SEO helps content be discovered, but a well-optimized page alone does not guarantee inclusion in AI answers. Dan’s perspective is that specific, credible customer experiences give businesses more useful evidence to offer, alongside sound search fundamentals.
For most of the last two decades, ranking well online meant one thing. Get the page right. Keyword density, header tags, backlink profiles, site speed, internal linking structure. An entire industry built itself around understanding exactly what a search algorithm rewards, then building pages that check those boxes as precisely as possible. If you tuned the page correctly, you had a real shot at page one.
That system worked because Google, at its core, was matching and ranking pages against each other. Your job was to make your page the best match for a given query and outrank whatever else was trying to answer the same question.
AI tools aren't playing the same game
When someone asks an AI tool a buying question, it isn't scanning for the best-tuned page and handing it back. It's pulling together an answer from whatever it can find, weighing which sources actually seem credible enough to inform that answer, and synthesizing something new rather than pointing to a single winning page.
That's a different kind of contest entirely. A page can be technically flawless, fast, well-structured, keyword-tight, and still contribute nothing to an AI-generated answer if the content underneath doesn't carry real evidence. Meanwhile a messier page, a forum post with typos in it, a customer story nobody ever ran through an SEO checklist, can end up shaping the answer if it happens to contain something genuinely specific and credible.
I've been watching this directly through the searches we track for ShareOne Reviews. None of the sources that keep showing up in AI answers are winning because someone tightened their metadata. They're winning because of history, specificity, and some signal of being real instead of written to perform.
What "ranking well" is starting to mean
Google ranks pages largely on signals about the page itself. Backlinks, structure, relevance, authority a domain accumulates over years. AI answers seem to weigh something closer to evidence quality. Does this source contain a specific, credible account of something that actually happened, or does it read like content built to rank rather than to inform.
Picture two pieces of content about the same product. One is a case study, three paragraphs, every sentence smooth, written entirely from the company's point of view, with a headline that hits the target keyword exactly. The other is a customer's own account, a little clumsy in places, mentioning a competitor they almost went with instead and a feature that annoyed them for the first two weeks before they got used to it. To a search engine built around page-level signals, the case study probably wins. To a system trying to judge whether something is credible, the messier account might actually carry more weight, because it doesn't read like it was built to persuade.
This doesn't mean traditional SEO stops mattering. Pages still need to be found, crawled, and indexed before any of this evidence weighing can happen at all. But that kind of technical groundwork is becoming table stakes rather than the thing that sets you apart. What sets you apart is whether there's something real sitting underneath it once someone, or something, actually reads it closely.
Where I think this leaves businesses
Most companies have spent years building processes around getting pages found and almost no process at all around generating real evidence. There's a content calendar for the blog and a checklist for every new page, but nothing equivalent for consistently capturing the specific, verifiable experiences that AI tools increasingly seem to weigh most heavily.
I think that's the gap worth paying attention to right now. Not because SEO stops mattering, but because it stops being enough by itself. A business with airtight technical fundamentals and no real evidence behind its claims is competing hard in a game that's quietly changing shape underneath it.
That's the shift Share One Reviews is built around, treating evidence generation as its own discipline instead of something bolted onto marketing after the fact. We're still early in figuring out exactly what these systems weigh most, and watching our own tracked searches move is turning into one of the more direct ways we're learning it, rather than guessing from the outside.