Who Feeds AI Answers for Supplier Queries?

When a buyer asks ChatGPT, Gemini, Claude or DeepSeek for suppliers, the answer draws on a limited source pool: B2B directories with structured company profiles, third-party list articles (often written by competitors from other countries) and, far less often, the manufacturers' own pages. The manufacturer itself is frequently absent from this chain; it is represented by other people's texts, whose errors it neither knows about nor can correct.

Three doors decide whether a company page works as a source: readability (content as HTML, not buried in PDF catalogues), extractability (short, machine-readable answers to concrete questions: capacity, tolerances, certificates, machine park) and credibility (one consistent company description everywhere plus mentions by third parties). An audit of eleven manufacturers (Gorilla76, 2026) adds a warning: self-praise lists ("best suppliers", written by the supplier itself) backfire; in 69 percent of cases the self-citing brand was excluded from the final recommendation.

The window is open: a Semrush analysis of 1,094 topic fields (2026) found a clear topic owner in only 15.2 percent of categories. Industrial sourcing topics largely belong to the unowned fields; whoever publishes solid, extractable first-hand information there first occupies the answer.