Sample report
A real AI visibility audit, published with its evidence
This is an audit we ran, unedited. We asked 29 buyer questions across ChatGPT, Perplexity and Gemini, several runs each, and logged all 232 answers along with every source the engines cited. Below is the opening of the report. The full document, including the evidence appendix, comes by email.
Page one: the score
A Adidas is the default answer
When buyers ask ChatGPT, Perplexity, Gemini or Google's AI for sportswear and athletic footwear in Europe, Adidas is named in 75% of answers; Nike in 44%. Named in 26 of 29 buyer questions (Nike: 15).
We deliberately published an audit of a brand that scores well. A disaster case would be easier to sell with and less honest about what most audits look like. Even at 87, the report finds three specific things costing the brand answers.
Page two: what we found
1 Asics appears in ChatGPT for 3 of the 3 “how to choose” questions; Adidas appears in 2.Buyers comparing options in ChatGPT are being sent to Asics by default.
2 The homepage was not reachable by our checker (HTTP 403); no explicit AI-crawler rules; no llms.txt.If automated visitors are blocked, engines cannot verify anything about the brand first-hand, so they fall back to competitors with cleaner records. Note this is what our checker saw — bot protection is the likely cause, and the report says so rather than claiming the engines are blocked.
3 When Adidas is named it sits at position #2.9 on average.The problem is being found, not being liked. Visibility fixes should convert quickly.
The evidence behind every number
Where the engines took their sources from
1,501 cited links across 232 answers, each classified. This chart appears in the report with the full source list.
Gemini100%
Perplexity98%
ChatGPT80%
All engines92%
Third-party pagesThe brand's own site
Get the full report
Ten pages: the score and how it is calculated, per-engine breakdown, all findings with fixes ranked by effort and impact, the earned-versus-paid action list, and the evidence appendix — every question, every answer, every cited source.
One email with the report. We will not add you to a list or pass your address to anyone. If you name a brand, we will tell you what we would look at for it — and we can run that audit free.
What is inside
- Score, grade and the formula behind it, with unmeasured parts excluded rather than assumed
- How each engine answers today, with pass / needs work / fail per engine
- Findings and fixes: what was found, why it matters, what to do, ranked by effort and expected impact across weeks 1–2, 3–6 and 7–12
- All fixes by angle, with earned work separated from paid placement
- Evidence A — every question, its estimated demand, per-engine result and who was named instead
- Evidence B — brands side by side, and where the answers come from
- Evidence C — brand record and website checks
- Method: how the prompts were built, how many runs, what was excluded and why
About this document. It analyses publicly available responses from AI answer engines, run on 18 August 2026. BrandsNode is not affiliated with, endorsed by, or commissioned by Adidas or any brand named in the report. The brand was chosen only because it is widely recognised. Competitor names appear because the engines named them in their own answers.