The Agent Selection Handbook
The handbook in reading order. Foundations explain how AI shopping agents choose products. The catalog chapters make your product data legible to those agents. The infrastructure chapters keep your store fast and open while agents read it.
Foundations
- Agentic Commerce OptimizationThe complete guide to agentic commerce optimization (AEO for ecommerce): make your catalog legible and preferable to AI shopping agents.
- How AI Agents Choose ProductsHow AI shopping agents rank products: the structured signals (price, availability, reviews, data completeness) that decide which store gets picked.
Catalog
- AI-Readable Product FeedAn AI-readable product feed leads titles with category and use-case and fills every attribute agents parse: the gaps are why agents skip you.
- Product Schema for AIProduct JSON-LD is the machine-readable layer agents trust when your page and feed disagree: ship Offer, price, availability, and review fields.
- Product Titles for AIAgents match shopper queries to titles almost literally. Lead with the words buyers type, not a brand name like 'The Luna'.
- Crawlers, robots.txt & llms.txtBlocking AI crawlers in robots.txt hides the page data agents read beyond your feed. Here's the allow-list and what llms.txt adds.
- Anatomy of a Citable, Comparable Product Detail PageWhat a product detail page needs so an AI agent can cite it and compare it: title, all-in price, availability, specs, reviews, and a legible substrate.
- Category and Collection Page Legibility for AI AgentsCategory and collection pages are how agents reach your products. Make them server-rendered, structured, and truthful so an agent can map your catalog.
Infrastructure
- Serving Agent TrafficHow to afford being open to AI agents: edge-cache the cheap machine surfaces, allowlist verified bots, and measure the load. Not block legit agents.
- Information DensityMaximize verifiable facts per token on machine surfaces. Same content, different format at one URL is negotiation; varying it by user-agent is cloaking.
- The Agent Selection Maturity ModelA five-level editorial framework for agent-selection readiness, from unreachable to measured, mapped to the selection checklist and signals.
- An Agent Selection Audit, Worked Through a Fictional StoreWalk the Agent Selection Checklist through a fictional store, category by category, to see what a real audit inspects and what to fix first.
Diagnostics
- Not in ChatGPT Shopping?Products missing from ChatGPT shopping? It's almost always data gaps: vague titles, missing availability, blocked crawlers. Diagnose in order.
- Shown but Not Selected: Why an Agent Picks a CompetitorYou appear in AI shopping answers but an agent keeps recommending a competitor. Diagnose which of four selection gaps is yours, then fix it.
- Lowest Price but Still Not Recommended by AI AgentsCheapest for the query but the AI agent recommends a pricier rival? Price is necessary but not sufficient. Diagnose why raw low price does not win selection.
- Why You Show in One AI Engine but Not AnotherShowing in ChatGPT but not Gemini or Perplexity for the same search? The cause is a rail, crawler, or data-source gap specific to the missing engine.
- When AI Shows the Wrong Price or Stock for Your ProductsYou appear in AI shopping answers, but an assistant quotes a stale price, an expired sale, or wrong stock. Diagnose which surface drifted, then fix it.