Chocolate,
according to AI
Every week we ask ChatGPT, Claude and Gemini the questions real UK consumers ask about chocolate — and record exactly which brands they put forward. Right now, Melt Chocolates owns the conversation.
0 AI answers this run · 0 brands named · measured, never estimated
Melt Chocolates is the default answer.
Named in 33% of chocolate answers, typically in position 1.8. Cadbury follows at 32%, Heroes at 32%.
Every brand AI names, ranked.
Visibility is the share of answers naming the brand; citation share is its slice of all brand-owned pages cited; read→cited is how often its pages convert a read into a citation.
| # | Brand | Visibility | Avg. position | Citation share | Read→cited |
|---|---|---|---|---|---|
| 1 | Melt Chocolates | 33% | 1.8 | 11% | 100% |
| 2 | Cadbury | 32% | 2.2 | 5% | 40% |
| 3 | Heroes | 32% | 2.2 | 0% | — |
| 4 | Galaxy | 20% | 2.9 | 0% | — |
| 5 | KitKat | 20% | 3.7 | 0% | 0% |
| 6 | Aero | 18% | 4.7 | 0% | — |
| 7 | Lindt | 17% | 3.2 | 5% | 0% |
| 8 | Hotel Chocolat | 15% | 2.7 | 20% | 21% |
| 9 | Maltesers | 15% | 3.4 | 0% | — |
| 10 | Quality Street | 15% | 4.4 | 0% | — |
| 11 | Milkybar | 15% | 5.3 | 0% | — |
| 12 | Twix | 13% | 3.6 | 0% | — |
| 13 | Milky Way | 13% | 3.8 | 0% | — |
| 14 | After Eight | 13% | 5.1 | 0% | — |
| 15 | Nestlé | 13% | 5.1 | 0% | 0% |
Melt Chocolates is named most. That is not the same as being wanted.
Being named is visibility. Underneath it, asked to rate each brand directly, the AI assistants judge it three ways — how much it stands out, how much it belongs, and how much it is returned to. They put 20% of the 55 brands rated above indifference. Bettys leads the brands the answers name, at 76% of maximum Desire. On attachment, the most divided driver, Bettys indexes 132 against Chocologic at 71, on a category average of 100.
Each whorl lights one ring per index threshold cleared (84 / 92 / 100 / 108 / 116). Above indifference means Desire above two-thirds of its maximum. See the full chocolate desire edition →
Before it answers, it goes looking for Hotel Chocolat.
An assistant does not answer from memory alone. It runs its own web searches first, and whoever it searches for is the shortlist it then chooses from. Across 76 searches in chocolate, Hotel Chocolat is named in 60% of the ones that name a brand at all. ChatGPT Web, Copilot do not report their searches, so these shares are measured against the ones that do.
These are the assistants' own searches, read from their answers — not what anyone typed into a search engine.
The referees of chocolate.
The independent sources AI models cite when answering chocolate questions — the places a brand needs to be seen, reviewed and ranked.
amazon.co.ukCadbury converts reads into citations. Hotel Chocolat doesn't.
When Claude or ChatGPT opens a brand's own pages while answering, does the page make the cited-sources list? Cadbury converts 40% of reads; Hotel Chocolat just 21%. That gap is a content problem, not a visibility problem.
Full data table
| # | Brand | Visibility | Avg. position | Citation share | Read→cited |
|---|---|---|---|---|---|
| 1 | Melt Chocolates | 33% | 1.8 | 11% | 100% |
| 2 | Cadbury | 32% | 2.2 | 5% | 40% |
| 3 | Heroes | 32% | 2.2 | 0% | — |
| 4 | Galaxy | 20% | 2.9 | 0% | — |
| 5 | KitKat | 20% | 3.7 | 0% | 0% |
| 6 | Aero | 18% | 4.7 | 0% | — |
| 7 | Lindt | 17% | 3.2 | 5% | 0% |
| 8 | Hotel Chocolat | 15% | 2.7 | 20% | 21% |
| 9 | Maltesers | 15% | 3.4 | 0% | — |
| 10 | Quality Street | 15% | 4.4 | 0% | — |
| 11 | Milkybar | 15% | 5.3 | 0% | — |
| 12 | Twix | 13% | 3.6 | 0% | — |
| 13 | Milky Way | 13% | 3.8 | 0% | — |
| 14 | After Eight | 13% | 5.1 | 0% | — |
| 15 | Nestlé | 13% | 5.1 | 0% | 0% |
The models do not agree on who leads.
Ask about chocolate and the brand you are shown depends on the assistant: Claude → Heroes, Gemini → Melt Chocolates, ChatGPT → Heroes. Their top-five lists share just 20% of names.
Brands whose visibility swings most between models — the bar per model, widest gap on the right.
The data behind this chart
| Brand | Claude | Gemini | ChatGPT | Spread |
|---|---|---|---|---|
| Melt Chocolates | 35% | 60% | 5% | 55pp |
| KitKat | 35% | 20% | 5% | 30pp |
| Milkybar | 30% | 15% | 0% | 30pp |
| Aero | 30% | 25% | 0% | 30pp |
| After Eight | 25% | 15% | 0% | 25pp |
| Nestlé | 25% | 15% | 0% | 25pp |
| Smarties | 25% | 15% | 0% | 25pp |
| Quality Street | 25% | 20% | 0% | 25pp |
| Lindt | 25% | 20% | 5% | 20pp |
| Ferrero Rocher | 20% | 0% | 0% | 20pp |
category-page carry chocolate.
Every page AI opened while answering, classified by what kind of page it is. category-page supply the most citations; Product pages convert best — 95% of the ones models opened were actually cited.
The data behind this chart
| Page type | Cited | Read but not cited | Conversion |
|---|---|---|---|
| category-page | 131 | 15 | 90% |
| Editorial | 123 | 78 | 61% |
| Product pages | 105 | 6 | 95% |
| How-to guides | 65 | 6 | 92% |
| Listicles | 48 | 12 | 80% |
| Reviews | 30 | 12 | 71% |
| Other | 28 | 57 | 33% |
| News | 18 | 6 | 75% |
Structured data is what separates cited from ignored.
Of the chocolate pages models actually opened, how much more often each feature's pages ended up in the cited sources. Structured data is worth 18 percentage points; cites its sources pages fare 64pp worse. Measured within the read population, so selection can't flatter it.
The data behind this chart
| Feature | Pages with | Cited % | Pages without | Cited % | Difference |
|---|---|---|---|---|---|
| Structured data | 570 | 78% | 273 | 60% | +18pp |
| Specifications | 155 | 83% | 688 | 70% | +14pp |
| Answer up front | 321 | 75% | 522 | 70% | +4pp |
| Named author | 328 | 68% | 515 | 75% | -7pp |
| Hard statistics | 348 | 63% | 495 | 78% | -15pp |
| Updated this year | 219 | 48% | 624 | 81% | -33pp |
| Comparison table | 159 | 45% | 684 | 78% | -33pp |
| Cites its sources | 84 | 14% | 759 | 79% | -64pp |
Kinder builds pages AI can quote.
Readiness scores each brand's own pages on the structure models reward — a direct answer up front, tables, hard numbers, named authors, structured data, recency, specifications. Melt Chocolates is the widest gap: well-built pages that still aren't converting into citations.
| Brand | Readiness | Pages cited | Pages profiled |
|---|---|---|---|
| Kinder | 72 | 75% | 4 |
| Diablo | 59 | 75% | 4 |
| Celebrations | 50 | 100% | 3 |
| Chococo | 43 | 60% | 5 |
| Firetree Chocolate | 38 | 75% | 8 |
| Galaxy | 32 | 33% | 9 |
| Harry Specters | 32 | 32% | 19 |
| Melt Chocolates | 30 | 7% | 135 |
| Lindt | 29 | 12% | 156 |
| Pump Street Chocolate | 29 | 15% | 20 |
Readiness is the share of eight measurable structural signals present on a brand's profiled pages. It is a build-quality measure, not a ranking of the brand.
The questions chocolate pages have to settle.
Read straight off the pages AI cited: the consumer questions cited content answers most often. Content that fails to settle these is content models have no reason to quote.
Nineteen more categories.
Method: 20 unbranded consumer prompts per category, spread across the buying journey, run weekly against ChatGPT, Claude and Gemini. Read→cited uses Claude and ChatGPT only (Gemini does not separate the two). Correlational observations of model behaviour at the time of the run. Data from the run completed Sat, 17 Oct 2026 13:10:34 GMT.