Protocol compliance
A valid /.well-known/ucp manifest with the right capabilities, payment handlers and signing keys. This is binary — either an agent can transact with you or it cannot.
Will ChatGPT recommend your store?
Paste your URL. In 60 seconds we tell you whether GPT, Claude, and Gemini will recommend you when a shopper asks about your category — then we hand you the three specific things you can fix today. Free, no account.
Try:
Probe
GET /.well-known/ucp, validate against schema, inspect services / capabilities / payment_handlers / signing_keys.
Score
Two structural layers — protocol checks (manifest hygiene) and content checks (Product JSON-LD, OG tags, identifiers, ratings, agent crawl headers).
Identify
Three frontier models read what an agent would discover from your page and tell you whether they could confidently identify and transact with your store.
Visibility
Same models, this time queried cold. Did your brand surface in their top picks? Cross-agent recall is the signal no single platform can give you.
What we actually measure
SEO trained an entire industry to think about keywords and backlinks. AEO runs on an almost opposite set of signals — agents weight what they can extract, quote and verify. We instrument every scan against the six levers that move LLM recommendation ranking, then run the same store through three frontier models to ground-truth it.
Quantitative density
Concrete numbers, specs, percentage claims and side-by-side comparisons. Pages that say '30% lighter, 287g, 4-season rated' outrank pages that say 'lightweight and durable'.
Quoted social proof
Verbatim customer quotes and expert reviews wrapped in semantic HTML or schema.org Review markup. Agents lift quotations directly into their answers.
External authority
Outbound links to manufacturers, certifications, press coverage. A page that cites trusted sources is preferentially recommended; a page that only links to itself looks like a brochure.
Structured data depth
Product JSON-LD with brand, GTIN/MPN, availability, aggregateRating. The richer the structured payload, the more confidently an agent can dedupe and rank your listing.
Cross-agent recall
The same category query asked to GPT, Claude and Gemini, cold. If two out of three forget you exist, the protocol and content layers don't matter yet — you have a brand-mention problem.
None of these are heuristics — every lever is measured against the structured signals an agent actually extracts from your page, then re-validated by running real queries through three frontier models on every scan.
SEO and AEO stack — they don't compete
Three audiences read your store. Your HTML body is shared across all of them — that's where SEO does its work and we don't touch it. The agent column adds four sidecar layers that SEO tools don't produce: a UCP manifest, an llms.txt index, agent-tuned fields inside your JSON-LD, and explicit AI-bot allow rules. That's the layer we optimize. What do agents actually read? →
Humans
- Rendered HTML body
- Marketing prose + brand voice
- Hero image, video, CTAs
Optimized by
Your designer + copywriter
Search engines
- HTML body (same as humans)
- Product JSON-LD
- sitemap.xml + robots.txt
- Open Graph tags
Optimized by
Your SEO tool
AI shopping agents
- HTML body (real-time browsing)
- Product JSON-LD
- + /.well-known/ucp manifest
- + llms.txt index
- + agent-tuned JSON-LD fields
- + AI-bot allow rules
Optimized by
aeoprepared.com (the + rows)
Why this matters
- → Your SEO investment is preserved. Body copy, internal links, keyword work — all untouched.
- → The four + rows above are sidecar surfaces that don't exist in standard SEO output. We add them; we don't edit what's already there.
- → No cloaking. Same HTML to humans, Googlebot, and AI browsing agents — agents get extra data they wouldn't synthesize otherwise.
- → Reversible. If you uninstall, your SEO surface is exactly where it was.
Common questions
What is Answer Engine Optimization (AEO)?+
Answer Engine Optimization is the discipline of making your commerce site discoverable, extractable and recommendable by AI shopping agents like ChatGPT, Claude, Gemini and Perplexity. Where SEO optimizes for keyword-matching crawlers, AEO optimizes for generative engines that quote, rank and recommend based on structured signals on the page plus brand mentions in their training corpora.
How is AEO different from SEO?+
SEO rewards keyword density, backlink count, and exact-match anchor text. AEO rewards the almost-opposite set of signals: extractable structured data, verbatim quotations the model can lift, concrete statistics and comparisons, outbound citations to authoritative sources, and brand mind-share in the training corpora. Keyword stuffing — a classic SEO move — actively hurts AEO ranking.
What is the Universal Commerce Protocol (UCP)?+
UCP is the open protocol AI shopping agents use to discover and transact with merchants. A merchant publishes a JSON manifest at /.well-known/ucp on their origin declaring services, capabilities (checkout, order, discount, fulfillment), payment handlers and cryptographic signing keys. Shopify ships UCP manifests by default for stores on the current platform tier; most merchants don't realize they already have one.
What does aeoprepared.com actually measure?+
Four layers on every scan. Protocol: a valid /.well-known/ucp manifest with the right capabilities, payment handlers and signing keys. Content: structured data, quantitative density, quoted social proof, external authority links, OpenGraph product tags. Identification: whether GPT, Claude and Gemini can correctly identify your product when handed your page directly. Visibility: whether the same three models recommend your brand in a cold-start category query with no link to your page.
What's the single highest-leverage AEO fix for most stores?+
Publishing a JWK array in the signing_keys field of /.well-known/ucp. In our benchmark of 27 top DTC stores, 81% publish a UCP manifest but 0% publish signing keys — the cryptographic primitive that lets a UCP agent verify the merchant's responses are authentic. It's three lines of JSON and immediately puts the store ahead of the entire DTC field.
Does aeoprepared work on non-Shopify stores?+
Yes. The audit is platform-agnostic — it fetches the URL, probes the origin's /.well-known/ucp endpoint, and parses structured data from the HTML. Shopify, WooCommerce, BigCommerce, Magento, Wix, Squarespace, headless commerce, and custom Next.js / Remix builds are all treated identically. The launch benchmark dataset is Shopify-heavy because that's where the highest DTC density is, not because the tool is Shopify-specific.
What do AI shopping agents actually read on my page?+
Three classes of agents read your store differently. Training-time crawlers (GPTBot, ClaudeBot, PerplexityBot) scrape HTML + JSON-LD and respect robots.txt. Inference-time browsers (ChatGPT browsing, Claude web search, Perplexity Sonar) fetch HTML live when a user asks a question and mostly ignore robots.txt because the request is treated as user-initiated. Commerce-specific agents (Shopify Agentic Commerce, Google Agentic Checkout, ChatGPT Shopping, Amazon Rufus) hit /.well-known/ucp first and use HTML as fallback. So HTML body matters for everyone; JSON-LD matters for everyone; UCP / llms.txt / AI-bot allow rules matter specifically for the agent surfaces SEO doesn't cover.
Will optimizing for AEO hurt my SEO?+
Not if you do it through the agent-only sidecar layers. Publishing /.well-known/ucp, llms.txt, AI-bot allow rules, and additional JSON-LD fields is purely additive — Googlebot doesn't penalize you for serving more structured data, the same HTML to all visitors, or honoring more crawlers. Where AEO can collide with SEO is body copy rewrites — if someone tells you to replace marketing prose with spec-sheet density, you risk losing keyword and brand-voice work that already converts. That's why our auto-fix prioritizes sidecar surfaces and treats body rewrites as opt-in, never automatic.
Why distinguish training-time crawlers from inference-time browsers?+
Because they obey different rules. Adding GPTBot to your robots.txt allowlist lets OpenAI's training crawler ingest your page into the next model snapshot — useful for long-term brand recall in ChatGPT. But when a user actively asks ChatGPT a shopping question and ChatGPT browses your page in real time, that traffic uses a different user agent and largely doesn't check robots.txt. The fix for inference visibility is structured data the live browser can parse fast — JSON-LD, OG tags, dense factual HTML — not crawler permissions.
Why AEO, why now
SEO optimized for crawlers reading HTML. AEO optimizes for agents answering shopping queries. The window to claim the category is now — before Shopify internalizes single-platform UCP checks, build the cross-agent visibility view (ChatGPT Shopping, Perplexity, Claude, Rufus, Gemini) that no platform has any reason to build for you.