THE ONE-PARAGRAPH VERSION
What answer engine optimization actually is
Answer engine optimization (AEO) — also called generative engine optimization (GEO) — is the practice of making a website structured and specific enough that an AI assistant can read it, trust it, and name it in an answer, the way search engine optimization once made a page rankable on a list of blue links. OpenAEO scores a site against 5 retrieval gates and 49 weighted checks across 6 dimensions, then returns the exact files and copy that move it above the ship floor. A failed gate — JavaScript-only rendering, AI bots blocked at the CDN, a robots.txt disallow, or unresolved placeholder text — caps the entire score, because a page an assistant cannot fetch or parse cannot be cited at all. Most sites clear the gates within two files and one rewritten paragraph. The audit is free on one site, needs no signup, and runs in about twenty seconds.
WHAT WE ACTUALLY SCORE
Five gates. Forty-nine checks. One honest number.
One audit covers both answer engine optimization (AEO) and generative engine optimization (GEO) — they're the same goal, measured across ChatGPT, Claude, Gemini and Perplexity. The gates come first, because they cap everything: a site can read as perfectly crawlable in robots.txt and still return 403 to every AI bot at the CDN. So we fetch your pages as each bot, from the edge, and check what actually comes back.
GATE 01
Renders without JS
Your text exists in the HTML a crawler receives, not after a bundle runs.
GATE 02
Bots reach the edge
Live per-bot fetch. CDNs and WAFs block AI crawlers by default — read, don't assume.
GATE 03
robots.txt allows
Every current AI agent, checked against a registry we keep updated, not a 2023 copy-paste.
GATE 04
Indexed where it counts
Several assistants lean on Bing's index. Not being in it is invisible with extra steps.
GATE 05
No placeholder text
Unresolved verification markers in live copy cap the score. Drafts don't get to pass.
ARetrieval & index
Can a machine get the page at all, and does it come back complete?
— static render / SSR diff
— per-bot edge response codes
— sitemap + llms.txt hygiene
— canonical correctness
BSchema
Structured data that mirrors what's visible — floating schema is the single most common failure we see.
— Organization + WebSite
— Product / Offer with price
— FAQPage matches visible Q&A
— truthful dateModified
CAnswer layer
Whether there is anything on the page an assistant can lift and quote without hedging.
— target answered in first 100 words
— self-contained passages, entity-led
— sequential H1 → H2 → H3
— visible freshness date
DPage coverage
One page per real buyer question — pricing, comparison, use case, proof.
— query → page mapping
— transparent pricing page
— comparison / alternatives page
— reviews with named sources
EOff-site consensus
Assistants trust agreement between independent sources more than anything you say about yourself.
— community mentions, earned only
— third-party roundup placement
— reviews + profile completeness
— NAP and entity graph match
FMeasurement
Whether you'd even know if it worked. Most sites can't tell.
— AI-referral traffic segmented
— weekly citation testing
— launch-day baseline captured
— quarterly re-audit scheduled
The score is the input. Citations are the outcome. We never hand over a scorecard without a same-day citation baseline beside it — a number you can optimize is not the same as a customer who found you.
Unverified facts get flagged, not guessed. Anything we can't confirm ships as a VERIFY marker in your copy and schema — and unresolved markers cap your score, so placeholder text can't ride a passing grade.
Off-site counts too. Assistants cross-reference. Community threads, third-party roundups, reviews and a consistent entity graph carry roughly a quarter of the score — because they carry roughly that much of the answer.
Questions people actually ask
What is AEO, in one sentence?+
Answer engine optimization: making your site structured and specific enough that an AI assistant can read it, trust it, and name you in an answer — the way SEO made you rankable on a page of links.
Is this just SEO with a new name?+
No. SEO competes for a ranked list. Assistants don't rank, they synthesize one answer from sources they can parse and verify. That rewards machine-readable structure, verifiable specifics, and consistent off-site signals — things classic SEO never scored you on.
What does the audit actually check?+
Five retrieval gates that cap your score when they fail — JS-only rendering, bots blocked at the CDN, robots.txt disallows, missing from the index assistants read, and unresolved placeholder text — plus 49 weighted checks across six dimensions: retrieval and index, schema, answer layer, page coverage, off-site consensus, and measurement.
How long until I get cited?+
It depends on crawl cadence, not on us. Most sites see their first citation change within 2–6 weeks of shipping the fixes. Remy's took 12 weeks to reach 9 of 12 prompts, with three flat weeks in the middle.
Do I need a developer?+
Not for the free tier — the fixes are two copy-paste files and rewritten page copy. Business plans output ticket-ready specs if you do have a dev; on Enterprise we implement.
Does it work on Squarespace, Shopify, Webflow?+
Yes. The audit reads your live HTML, so the platform doesn't matter. The instructions are written per-platform, including exactly where to paste code in Squarespace and Shopify.
Can you guarantee I'll show up in ChatGPT?+
No, and be suspicious of anyone who does. We can guarantee you clear the technical bar that's keeping you out, and we measure whether you're being cited — so you're never guessing.
How much does OpenAEO cost?+
The audit is free forever on one site, with no signup and no card. Solo is $10 per month for one site with monthly re-crawls and an AI agent that edits your uploaded source directly. Business is $39 per month for 5 sites with weekly crawls and citation tracking across four assistants. Enterprise is custom. See
pricing.
What counts as a passing AEO score?+
Scores run 0 to 100. The build ship floor is 72. Below 46 a site is barely retrievable; 46–64 is AI-ready, 65–82 is AI-competitive, and 83 or above is AI-dominant. A failed retrieval gate caps the score no matter how strong the rest of the page is.
More questions? See the full FAQ.
