When AI is asked about coffee — who does it recommend?

When AI recommends coffee, who actually gets named?

Ask ChatGPT or Google's AI Overviews which coffee to buy, and you won't get a list of ten links to scroll. You'll get a sentence or two that names a couple of brands — and ignores everyone else. We measured exactly which brands make that cut.

The method is deliberately literal. We sent 24 real buyer questions about coffee to 2 AI engines — ChatGPT and Google AI Overviews — and recorded every brand the models named. Across those answers AI mentioned 108 distinct providers a total of 301 times. A brand's Share of Voice is its slice of those 301 mentions: how often AI brings it up relative to everyone else. It is not a measure of sentiment — a brand named as a cautionary example still counts — but it is the closest proxy we have for the question every marketer is now asking: when AI answers, does it say my name?

This is a different game from classic SEO. Ranking #4 on Google still puts you on page one, where a curious buyer can find you. There is no page one in an AI answer. The model picks its shortlist, states it with confidence, and moves on. Brands that are named are effectively the only options the user is shown; everyone else is invisible at the precise moment of decision — not ranked lower, but absent from the conversation entirely.

That is why share of voice in AI answers behaves less like search rankings and more like brand recall in a crowded room: a few names get repeated, and repetition compounds. The data below shows just how lopsided that dynamic already is for coffee.

AI's coffee shortlist is far shorter than the shelf

Start with the headline number. AI named 108 different coffee brands across the 24 prompts — a genuinely long list. Yet Lavazza alone accounts for 9.3% of every brand mention (28 of 301), and Tchibo a further 9.0%. Together the top three take 25.6% of the entire conversation.

The concentration is the real story. The top six brands capture 36.9% of all mentions — roughly half — while the remaining 96 brands split what's left. That long tail — 96 brands — shares just 50.8% between them, averaging a fraction of a percent each. In practice, being the 30th-most-named coffee brand and being unnamed are the same outcome: invisible.

The drop-off is steep, not gradual. Lavazza is named about 1.3× as often as the brand in third place (Melitta, 7.3%), and the curve keeps falling fast after that. This is not a tight race with a clear runner-up — it is a winner-take-most market where a couple of names absorb most of AI's attention and everyone else competes for scraps of recognition.

A few names, then a cliff: the top six coffee brands take ~36.9% of every AI mention — 96 brands share the remainder.

How to read the ranking above: each bar is a brand's share of all 301 brand mentions across the 24 prompts. Longer bar = named more often by AI.

The 96+ coffee brands AI almost never names

It is worth dwelling on the brands you don't see. Beyond the names above, AI also mentioned roughly 96 further coffee brands — but together they account for only 50.8% of all mentions. Spread across that many names, the typical also-ran is recommended in well under one prompt in a hundred. For those brands, the practical AI visibility is effectively zero: a buyer running these questions would almost never see them.

This is the part that doesn't show up in a normal analytics dashboard. A brand can have a healthy website, real customers and steady search traffic, and still be functionally absent from the AI layer that an increasing share of buyers now consult first. The long tail isn't a ranking problem to optimise — it's an awareness problem the brand may not even know it has, because nothing on its own site reveals it.

Why being named compounds — and being absent costs

For a brand already in the top few, this is a moat. Every time AI repeats Lavazza as the answer for coffee, it reinforces the model's prior that Lavazza is the answer — the same compounding loop that makes a market leader harder to dislodge with each mention. AI doesn't just reflect brand awareness; it amplifies it.

For everyone else, the takeaway is uncomfortable but clear. You are not competing for a higher position — you are competing to be mentioned at all. That means the levers are different from classic SEO: it is less about a single page ranking and more about the breadth and consistency of how your brand, products and category are described across the web that these models train and ground on. Brands that show up in comparisons, structured product data, reviews and authoritative third-party sources give AI more reasons to name them.

The first step, though, is simply knowing where you stand — which prompts you appear in, which you don't, and who AI names instead. That is exactly what this report measures, and what BuzzView tracks continuously across 8+ AI engines.