Love that you’re pushing merchants to treat llms.txt like a public-facing spec sheet — we’ve had better luck when it mirrors our product taxonomy and priority FAQs instead of just a blanket allowlist. Chunkable PDP copy has been huge for us too; rewriting features into single-question paragraphs made agents cite us far more often during browsing sessions. Another lever has been syncing the same structured data into Merchant Center and Perplexity Collections so the models can corroborate price and availability without guessing.
Great additions @Geoffy the Merchant Center and Perplexity Collections is something merchants are beginning to think about. The key is to build AI trust signals by comparing prices and availability.The spec sheet understanding for llms.txt is spot on; it’s less about crawling permissions and more about giving AI a structured brief about your store. Are you seeing any specific product categories where chunkable PDP copy moves the needle most?
Worth noting that Shopify activated Agentic Storefronts by default for all eligible stores on March 24 — so many merchants are already discoverable in ChatGPT without doing anything extra. Hence, one more real question to ask is: how do you actually know if your store made the cut, and whether it’s already generating AI-driven orders (lucky you if so)?
Rather than guessing, Stabile’s AI channel check measures actual AI-attributed order activity, so merchants know for sure what’s happening — not just whether the channel is theoretically enabled or not.
