AI answer engines are the new homepage: optimise for mentions, not rankings

by Maya Patel

Ten years in technical SEO and 2026 is the first year my old playbook actively misleads clients. A growing slice of buyers never see ten blue links now - they ask an assistant and get one synthesised answer with three or four sources baked in. If you're not cited, you don't exist, and there's no page 2 to crawl back from. What that changes: 1) Mentions beat pages - the model cites entities corroborated across trustworthy places, not whoever stuffed the most keywords. PR and SEO have merged. 2) One canonical explainer beats forty thin posts. 3) No more "ranking" to watch - only a test: ask the top assistants your buyer's real questions, see if you're in the answer. Anyone auditing their "am I in the answer" rate yet? I suspect B2B is well ahead of consumer.

favorite 32 comment 12 visibility 641

Comments

James Carter 1 month ago

This lines up with what I see on the content side, Maya. The uncomfortable part: the winning asset isn't a better page, it's a mention you don't control. A recent read had branded web mentions ~3x more predictive of LLM citations than backlinks - so half the job is now PR, podcasts, being the name quoted in someone else's explainer. On B2B being ahead: agreed, but it's partly survivorship. B2B has fewer buyers asking narrower questions, so "am I in the answer" is measurable. Consumer is a thousand fuzzy queries, and the same study put ~90% of brands at zero AI mentions. Depth consolidates; volume dilutes you across engines that each cite differently. Nobody shares their bank's blog - but an assistant quotes it if enough trustworthy places already do.

Olivia Chen 1 month ago

Builder note: I actually ran this audit last month for our dev tool. Took our real buyer questions ("best X for Y", "X vs Z"), fed them to the top three assistants, logged whether we showed up. Two lessons. One, it isn't one leaderboard - the same query cited almost totally different sources per engine (I've seen the "only ~2% of citations overlap all three" number and yeah, that tracks). So you're not optimising for a ranking, you're auditing 3+ separate judges. Two, the thing that moved the needle wasn't more posts, it was getting our comparison page corroborated on sites we don't own. B2B ahead? For us the audit was easy precisely because the buyer questions are finite. I just wish I'd started tracking it as a metric a year ago.

Priya Nair 1 month ago

Data-person reflex: if you can't measure your citation rate, you don't have a channel, you have a vibe. The trap is treating "am I in the answer" as yes/no - it's a distribution. Same question, run 20 times across engines, and you get a citation frequency with real variance, not a single truth. Track the rate over weeks, not the one screenshot where you happened to show up. Otherwise you're celebrating noise.

Will Blanc 1 month ago

Maya, one wrinkle from the byline side: being cited and being mentioned are different products, and most people are buying the wrong one. Roughly 62% of AI citations are "ghost" ones - the model links the source and never says the brand name out loud. And the engines are mirror images. Gemini names a brand in ~84% of appearances but links ~21% of the time; ChatGPT does the reverse, ~87% cite, ~21% mention. So Priya is right that it is a distribution, but it is two distributions and they trade against each other. Olivia, your comparison page moving the needle tracks - "X vs Y" pulls ~1.9 citations per retrieval, ~45% above average. I write those explainers for a living. Now the assistant quotes my paragraph and names nobody.

Maya Patel 1 month ago

Will - "now the assistant quotes my paragraph and names nobody" is going in my next client deck, with attribution, because apparently that's a thing I have to demonstrate out loud now. The ghost split genuinely reframes my point 3 and I'll own it: "am I in the answer" was too coarse. Two columns, not one - cited, and named - because what clients are actually buying is the second and I've been quietly selling them the first. Priya makes it three: run it twenty times, take the rate. To my own question, since it's rude to ask and not answer: across my audits this year, B2B lands somewhere in the answer maybe 40% of the time. Consumer is close to nothing, matching the ~90%-at-zero figure James quoted. Ahead, but mostly because the questions are countable.

Rosa Engel 1 month ago

This community is quickly becoming my favourite corner of the internet.

Kofi Papadopoulos 1 month ago

The shift nobody has priced in: answer engines cite sources they trust, and trust accrues to entities, not pages. The winning move is one canonical, ruthlessly maintained explainer per question you want to own — depth consolidates, volume dilutes.

Kofi Papadopoulos 1 month ago

Maya, let me add the plumbing under your two columns, since that half is my job. Being NAMED is downstream of being an entity. A model names a thing it can resolve: one canonical name, one description, corroborated in enough independent places that the graph collapses your forty variants into a single node. Most audits I run fail there before they fail on content. Brand name, product name and legal entity all differ, three spellings loose in the wild, no sameAs anywhere. The model ends up holding three fuzzy things, can't safely name any of them, so it links instead. That is Will's ghost citation, seen from the pipes. Cheapest win I sell all year: pick one name, use it everywhere, get it corroborated off your own domain. Deeply unsexy. Moves the named column.

Farah Castillo 1 month ago

The SEO threads here are better than most paid newsletters.

Naomi Moreau 1 month ago

Watching my traffic mix shift month by month: classic search slowly down, AI referrals up from zero to 4%. Small but the slope is the story.

Eitan Gonzalez 1 month ago

Digital PR and SEO are the same job now. The sooner clients accept that, the sooner the numbers move.

Yusuf Schneider 1 month ago

Structured data went from nice-to-have to table stakes. The machines are the audience now.