There’s a test worth running before any conversation about AI visibility, because it takes ninety seconds and it replaces speculation with evidence.
Open ChatGPT. Ask: “What’s the best [your category] brand?” Then refine: “I’m looking for [specific use case].” Then go specific: “For [exact spec], which brand do you recommend?”
That’s three depths of buyer intent — Broad, Mid, Ultra. We call them Intent Tiers. Between them, those three queries cover almost every shopping conversation a buyer has with an AI assistant, and they map almost cleanly onto the funnel ecommerce teams already understand.
Run it honestly — no brand name in the query — and the results are frequently sobering. Real brands doing seven and eight figures, brands that rank first on Google for their core terms, routinely miss at the Mid tier. Many miss at Broad too. That isn’t a corner case. It’s the normal state of a market that optimised for one judge for twenty years and is now being evaluated by another.
Why this happens
It isn’t that AI dislikes these brands. Assistants work on a coherence model — the sources they pull from, the structured signals they can read, the cross-references they can match. If a brand’s product pages don’t speak that language, the assistant has nothing to surface.
Most ecommerce sites were built for Google. Google rewards keywords, backlinks, page authority. AI rewards machine-readable claims, verifiable specs, and consistent brand identity across the web. Different game, different ground.
The result: a brand that ranks first on Google for “cycling bib shorts” can be invisible when someone asks ChatGPT for cycling bib shorts. That’s not the brand being punished. That’s the brand showing up in the wrong room.
Reading your result
The tier where you miss tells you which conversation to have internally:
- Miss at Broad — a structural problem. The assistant can’t establish who you are or that you belong in the category at all. This usually means missing or contradictory foundational data across your site and the sources AI reads.
- Miss at Mid — a content problem. You exist, but nothing connects your products to the use cases buyers actually describe. Comparison and use-case content, structured properly, is the lever.
- Miss at Ultra — a specification problem. Your product data doesn’t carry the precise, machine-readable attributes that let an assistant match you to an exact requirement.
Three different conversations, all starting from the same three queries.
Why acting early compounds
The next twelve months will separate brands that worked this out early from ones that didn’t. Not because AI is replacing search wholesale — at the high-recall end of the funnel, brand strength still matters — but because brands that show up consistently across Google, ChatGPT, Perplexity, Gemini and the agentic layer behind them compound in one direction, and brands missing from two of the five compound in the other.
Most ecommerce leaders know this directionally. What we see in conversations is hesitation about where to start, and pessimism about whether it’s worth investing before the gap shows up on a dashboard. The trouble is that this channel doesn’t have a dashboard — the loss is silent. The ninety-second test is the cheapest instrumentation you’ll ever install.
Take the ninety seconds today. The result tells you which conversation you should be having — and how soon. When you’re ready to go deeper than three queries, Geoffy’s approach to structured product data picks up exactly where the test leaves off.
Prefer a full diagnostic across engines and intent tiers? Get your free GEO Score.