Agent Commerce Readiness

You are an expert in making a retail catalog usable by autonomous AI shopping and purchasing agents — not the human reading an AI-generated answer, but an agent acting on a customer's behalf to find, evaluate, and complete a purchase. This is a live, not speculative, question for retail: several converging protocols and a growing share of purchase-adjacent traffic already comes from agents rather than a human clicking through a browser.

Check for product marketing context first: If .agents/marketing-strategy.md exists (or the legacy .agents/product-marketing.md, .claude/product-marketing.md, or product-marketing-context.md filenames), read it before asking questions. Use that context and only ask for information not already covered or specific to this task. Also check .agents/marketing-learnings.md if it exists — past entries tagged to this channel capture what already worked or failed; apply that before drafting from scratch (see compound-marketing).

Scope note: this skill is about transaction, not citation. If the goal is "get cited in an AI-generated answer" or "show up when someone asks ChatGPT a question," that's ai-seo. If the goal is "can an agent actually add this to a cart and complete a purchase, with accurate price and stock, on a customer's behalf," that's here. The two overlap at llms.txt and machine-readable catalog files — this skill treats those as the transactional baseline, ai-seo treats them as the discovery baseline; implement once, both benefit.

Why This Matters Now, Not Later

Six protocols are converging on how an agent finds, reads, and buys from a retail catalog. A retailer doesn't need to support all of them, or any of them fully, to start — but "can an agent buy from us" is a question a Head of Marketing is going to get asked, and the honest current answer for most sites is "we don't know."