Field research · Auckland · June–July 2026

Every vertical has a
kingmaker platform.

Ask the engines about barbers and one booking platform dominates the citations. Ask about plumbers and it's a tradie-review site. Ask about restaurants and the booking layers appear again and again. In every vertical we measured, there's a platform quietly deciding who's on the AI shelf.

SCOPE — Auckland, New Zealand · June–July 2026 snapshot. 7 verticals, 679 probed businesses: Italian, Japanese, sushi and Indian restaurants, barbers, plumbers, and board-game shops. Engines queried via their APIs — Gemini with Google Search grounding, OpenAI web_search, Claude with live search, and Apple Maps (MKLocalSearch, the layer Siri uses for local queries). APIs are clean, reproducible instruments and proxies for the consumer products, not the apps themselves. One city, one snapshot — treat this as field data, not a global law.

When we pulled apart where the engines learned about the businesses they recommended, a pattern repeated in every vertical: alongside the editorial lists and Reddit threads, each trade has one or two platforms that function as kingmakers — be on them and the engines can see you; skip them and you're relying on someone writing an article about you.

  • Barbers → Fresha. The booking platform drew 34 engine citations in our barber run alone — for many shops, their Fresha profile is their machine-readable presence, whether they think of it that way or not.
  • Restaurants → the booking layer. OpenTable and FirstTable listings recur across our restaurant runs as the places engines verify that a restaurant is real, bookable, and priced.
  • Plumbers → NoCowboys. New Zealand's tradie-review site keeps appearing as the trust source for a vertical with almost no editorial coverage — nobody writes listicles about drain repair.
  • Retail → the community's own institutions. For board-game shops it was Heart of the City and BoardGameGeek. Specialist communities document their shops better than any marketing department could.

The common thread: engines want a source that is structured, maintained, and independent-ish. A vertical platform is all three. Your own website is only one of them.

Two honesty notes before you act on this. The per-platform citation counts come from one run per vertical, and single-run counts wobble (measurably). And the specific names are New Zealand's — NoCowboys and FirstTable won't exist in your country, though something is almost certainly playing the same role. "Be on the platform" is also correlation-shaped advice: we haven't yet watched a business join one and start surfacing. What we observed is that the surfaced businesses were overwhelmingly the platformed ones.

The uncomfortable part

This means part of your AI visibility is delegated to a platform you may treat as an afterthought — a booking profile someone set up in 2023, a review page you've never claimed. In our runs, engines cited these platform pages instead of businesses' own sites routinely. If the platform's version of you is stale, that's the version being recommended. (It gets stranger: some platforms are now speaking to AI agents on your behalf without you knowing — we verified a case of that in Your platform is already talking to agents.)

What to do about it

Find your vertical's kingmaker and treat it as a first-class channel. Ask an AI assistant for recommendations in your category and look at which platforms it cites — that list is your priority order.

Then audit what those platforms say about you. Hours, pricing, services, photos — on the platform, not just your site. In our data, the platform profile was frequently the thing actually being read.

Who's kingmaking your vertical?

We map which sources drive AI recommendations in your category, and what they currently say about you.

Request a shelf check →