# Agentic Readiness Checklist > The manual version of the Agentic Readiness Audit. Ten categories, in the order the automated scan weights them. Work top-down: the categories at the top of this file move the score most because they decide whether an agent can *act*, not just read. > > Full scoring method: https://agenticstorefront.com/resources/readiness-methodology/ > Sample report from a real scan: https://agenticstorefront.com/resources/sample-readiness-report/ > Run it automatically (free, no account): https://agenticstorefront.com/assessment/ > > Agentic Storefront by ForkPoint · MIT-licensed UCP cartridge for SFCC: https://github.com/ForkPoint/ucp-sfcc Scoring reminder: each category is scored on its applicable checks only, then multiplied by the fixed weight below. Weights sum to 1.0. Tiers: 90+ Agent Ready, 70–89 Partially Ready, 50–69 Needs Work, below 50 Not Ready. --- ## 1. AI Agent Tools & Action Surfaces — weight 0.18 The only category that decides whether an agent can transact rather than merely read. - [ ] An agent-readable catalog or capability manifest is published and returns 200 to non-browser clients - [ ] Product, search, and cart actions are reachable without executing your front-end JavaScript - [ ] Checkout is exposed as a callable surface, not only as a click path - [ ] Declarative forms carry names, types, and required flags an agent can map - [ ] Tool or endpoint descriptions state what each action does and what it returns - [ ] Error responses are structured and machine-parseable, not HTML error pages - [ ] Rate limits and auth requirements for agent traffic are documented publicly ## 2. Content Discoverability — weight 0.15 Whether an agent can find and fetch your content at all. - [ ] `/llms.txt` exists at the site root, with an H1, a blockquote summary, and H2 sections - [ ] Every link in `llms.txt` resolves and carries a description - [ ] `sitemap.xml` exists, is current, and is referenced from `robots.txt` - [ ] Product and category content is present in the server response, not injected after hydration - [ ] Canonical URLs are set and self-consistent - [ ] Pagination exposes real URLs rather than infinite scroll only - [ ] Product feeds are reachable at a stable URL ## 3. Structured Data & Schema Markup — weight 0.12 - [ ] At least one valid JSON-LD block on every template - [ ] Every block has `@context` and a valid `@type` - [ ] `Organization` schema with name, url, logo, and `sameAs` - [ ] `WebSite` schema with a `SearchAction` using a `{search_term_string}` template - [ ] `Product` schema with an `Offer` node: `price`, `priceCurrency`, `availability` - [ ] `BreadcrumbList` on category and product templates - [ ] `FAQPage` on pages that actually answer questions - [ ] Review and rating markup reflects real, on-page reviews ## 4. Technical Readiness & Security — weight 0.09 - [ ] Agent user-agents receive 200s, not challenge pages or CAPTCHAs - [ ] Bot mitigation allowlists named AI crawlers and agent traffic - [ ] TLS valid, no mixed content, HSTS present - [ ] Redirect chains are one hop and preserve the path - [ ] Response times stay under a couple of seconds for agent-facing endpoints - [ ] Rate limiting returns `429` with `Retry-After`, not a silent block ## 5. AI Crawler Permissions — weight 0.08 - [ ] `robots.txt` explicitly names GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and Applebot-Extended - [ ] No blanket `Disallow: /` that silently removes you from the answer layer - [ ] `X-Robots-Tag` headers agree with `robots.txt` - [ ] The decision to allow or block each crawler is deliberate and documented - [ ] CDN or WAF rules do not contradict `robots.txt` ## 6. Meta Tags & AI Head Elements — weight 0.08 - [ ] Unique, specific `` per page — no templated repetition across the catalogue - [ ] Meta description present and written for a person, not stuffed - [ ] Open Graph and Twitter cards complete on shareable templates - [ ] `lang` and locale signals correct on every template - [ ] Product meta (price, availability, brand) present where applicable ## 7. Semantic HTML & Content Structure — weight 0.08 - [ ] One `<h1>` per page, headings nested in order - [ ] Landmarks present: `header`, `nav`, `main`, `footer` - [ ] Price, availability, and specs live in text nodes — never in images or CSS backgrounds - [ ] Specs and comparisons use real `<table>` or list markup - [ ] Key commerce facts are not hidden behind tabs that never render server-side ## 8. Generative Engine Optimization — weight 0.08 - [ ] Each key page carries a definition or claim a model can lift verbatim - [ ] Entity naming is identical across the site, schema, and off-site profiles - [ ] Comparison content exists and concedes honestly where a competitor wins - [ ] Claims carry sources, dates, or figures — not adjectives - [ ] Methodology behind any published score or framework is public ## 9. Accessibility & Agent Interaction — weight 0.07 The same signals an agent uses to operate a page. - [ ] Every input has a programmatic label - [ ] Buttons and links have accessible names that describe the action - [ ] Focus order follows visual order; no keyboard traps - [ ] ARIA roles are valid and not contradicting native semantics - [ ] Images that carry information have meaningful `alt` text ## 10. Answer Engine Optimization — weight 0.07 - [ ] Question-shaped headings that match how buyers actually ask - [ ] A direct answer in the first paragraph under each question - [ ] FAQ content marked up and mirrored in `FAQPage` schema - [ ] Answers are self-contained, so a quoted paragraph still makes sense - [ ] Freshness dates on anything time-sensitive --- ## What this checklist cannot cover The automated scan and this checklist both read your public storefront. Four of the seven readiness dimensions need access you would have to grant: - **Data Quality** — full-catalogue attribute completeness, inventory sync, pricing consistency - **Transaction Enablement** — payment handler exposure, session security - **Post-Purchase Experience** — order status, tracking, and returns APIs behind auth - **Measurement & Optimization** — agent session attribution in your analytics stack Those are what the 126-point expert assessment adds. It is free and delivered in three business days: https://agenticstorefront.com/services/readiness-assessment/