Visibility in AI Search: What Makes ChatGPT Recommend My Brand? (DMEXCO 2026)
Summary
Peec AI published a recorded DMEXCO masterclass in which Malte Landwehr presents what the company's brand tracking data shows about how ChatGPT picks sources after its latest model update. The talk reports that listicles, Reddit threads, and comparison pages are losing ground as citation sources, while pricing pages, documentation, PDFs, and.gov and.edu domains are gaining.
The second half is tactical. It covers reading fan-out queries to find the category terms a model searches for, breaking content into short, self-contained, fact-dense blocks, publishing pricing comparison pages about competitors, and using a plain markdown page of declarative facts as a citation target.
Demonstration by Peec AI, embedded from YouTube. AIOTruth did not reproduce the test.
Why it matters
If the pattern Peec AI describes holds, the pages a brand relied on for AI citations are no longer the ones being pulled into answers. Someone trying to get recommended by ChatGPT would shift effort away from listicle placements and Reddit mentions and toward the pages a model retrieves when it runs site-specific searches: pricing, documentation, and plainly stated facts that survive being quoted out of context. The fan-out query method also gives a concrete way to learn which category terms a model actually searches for, instead of guessing.
Source
- peec.ai/ the item itself, published by Peec AI, created by Peec AI
- peec.ai/changelog Reproducible first-party demonstration, first party
- docs.peec.ai/intro-to-peec-ai Named expert analysis, or a first-party claim about itself, first party
- youtube.com/watch?v=zUfwkh3eybMwhere AIOTruth found it
What was checked
- canonical URL taken from the page itself
- read the item page (200)
- https://docs.peec.ai/intro-to-peec-ai: supports the item, tier 8 (Named expert analysis, or a first-party claim about itself)
- https://peec.ai/changelog: supports the item, tier 2 (Reproducible first-party demonstration)
- https://peec.ai/: supports the item, tier 8 (Named expert analysis, or a first-party claim about itself)
Limitations
- Everything checked traces back to a single domain, so this is one party's account.
- The demonstration was not independently reproduced by AIOTruth.
AIOTruth judgment
Relevance is at the top of the scale because the talk addresses how a brand gets cited and recommended by ChatGPT, which is the core subject here. Evidence scored well and the status is Supported because the claims come from a named speaker drawing on the publisher's own tracking data, and the publisher's site, documentation, and changelog confirm the product and its data exist. Originality is solid because the findings are first-party observations instead of a recap of other reporting. Usefulness is the weakest dimension: the available material is a description and chapter list, so the tactics are named but not laid out in enough detail to apply without watching. Everything checked traces back to a single domain, so this is one party's account, and the demonstration was not independently reproduced by AIOTruth.
How this score was calculated
| Dimension | Weight | Score | What produced it |
|---|---|---|---|
| AIO relevance | 30% | 10 | ai search, geo, aeo, llm visibility, ai mode; 8 scope question(s) matched |
| Usefulness | 30% | 5 | 0 artifact(s), 4 actionable marker(s), 1 measured figure(s) |
| Evidence | 25% | 8 | Reproducible first-party demonstration; 3 verified source(s) across 1 domain(s) |
| Originality | 15% | 7 | no original testing found; closest archive match 0.19 |
aioRelevance x 0.30 + usefulness x 0.30 + evidence x 0.25 + originality x 0.15. The rubric is published in full on the Editorial Method page. Scoring is deterministic: the same item scores the same on every run.
Watch the demonstration Open the source
Found something wrong here? AIOTruth corrects material errors openly. Have something we should review? Submit a find.