AI-Discoverability Checklist
Can Google, ChatGPT, Claude, and Perplexity actually find and understand your website? Seven checks, one scoresheet, plain answers.
Why this matters now
People don't only search on Google anymore. They ask ChatGPT, Claude, and Perplexity questions and expect a direct answer with a source. If your website isn't set up for both traditional search and AI tools to read it, you're invisible in a growing share of how people actually look for businesses like yours.
This is the same 7-point standard we hold every JalenBuilds Studio launch to before calling a site "done." It's not a vague best-practices list — it's the exact checklist, in the exact order we run it.
How to use this
Go through the 7 checks one at a time. For each, mark whether it's done, partially done, or missing. Add up your score at the end. Most sites score lower than they expect — that's normal, and every gap here has a plain-English fix.
The 7-point standard
Google Search Console — verified
Why it matters: This is how Google knows your site exists and lets you see how people find it. Without it, you're invisible to Google's own tools and slower to get indexed.
How to check: Go to search.google.com/search-console. Is your domain listed as a verified property? If you don't have an account, it's not set up.
How to fix: Add the property, verify with the HTML tag method (your web developer adds one line of code to the site), then submit your sitemap inside Search Console.
Sitemap — present and submitted
Why it matters: A sitemap is a list of every page on your site, handed directly to search engines. Without one, search engines have to guess which pages exist by following links — and they miss things.
How to check: Visit yoursite.com/sitemap.xml. Does a page load listing your URLs? Has it been submitted inside Google Search Console?
How to fix: Have your developer generate a sitemap listing every public page, then submit the URL inside Search Console.
robots.txt with AI crawler rules
Why it matters: This file tells search engines and AI tools which pages they're allowed to read. If it's missing or blocks the wrong things, AI tools like ChatGPT's browsing feature or Claude may not be able to read your site at all.
How to check: Visit yoursite.com/robots.txt. Does it load? Does it explicitly allow crawlers like GPTBot, ClaudeBot, and PerplexityBot — not just block everything?
How to fix: Add a robots.txt that allows your public pages for both search engines and named AI crawlers, and points to your sitemap.
llms.txt — an AI-readable summary
Why it matters: This is a plain-text file written for AI tools specifically — a short, direct summary of what your business does and where the important pages are. It's the newest of the seven checks and the one most sites are missing entirely.
How to check: Visit yoursite.com/llms.txt. Does anything load?
How to fix: Add a short text file: one line describing your business, 2-3 sentences of context, and links to your most important pages.
JSON-LD structured data
Why it matters: This is hidden code that tells search engines exactly what kind of thing your page is — a business, a product, an article — plus specific facts like your name, address, and hours. It's how Google shows rich results (star ratings, prices, hours) directly in search.
How to check: Paste your homepage URL into Google's Rich Results Test. Does it detect any structured data?
How to fix: Have your developer add basic Organization and WebSite structured data to your homepage, at minimum.
Social preview tags (OG + Twitter cards)
Why it matters: When someone shares your link in a text, on Facebook, or on X, these tags control whether it shows a clean preview card with a title, description, and image — or just a bare, ugly link.
How to check: Paste your site's link into a text message to yourself, or use opengraph.xyz. Does a preview image and description appear?
How to fix: Add Open Graph and Twitter Card meta tags to every page — title, description, and a preview image sized correctly for social platforms.
Canonical URLs
Why it matters: If your site is reachable at multiple URLs (with and without "www," with and without a trailing slash, an old staging link), search engines can split credit between them or get confused about which is the "real" page.
How to check: View your page source (right-click, "View Page Source") and search for rel="canonical". Is there exactly one, pointing to your real domain?
How to fix: Add a canonical tag to every page pointing to its one true URL on your production domain.
Your scoresheet
Mark each item: 2 points if fully done, 1 point if partially done, 0 if missing. Add them up out of 14.
| Check | Done (2) | Partial (1) | Missing (0) |
|---|---|---|---|
| 1. Google Search Console verified | |||
| 2. Sitemap present + submitted | |||
| 3. robots.txt with AI crawler rules | |||
| 4. llms.txt present | |||
| 5. JSON-LD structured data | |||
| 6. OG / Twitter social tags | |||
| 7. Canonical URLs |
What your score means
12-14 points — Fully discoverable
Google and AI tools can find, read, and understand your site. Keep it maintained — sitemaps and structured data can drift out of date after big content changes.
6-11 points — Partially visible
Search engines can probably find you, but AI tools may be missing key context, and you're likely losing rich-result opportunities in Google. Fix the 0-point items first — they're usually the fastest wins.
0-5 points — Largely invisible
Your site is likely relying on people typing your exact business name to find you. Organic discovery — through search or AI tools — is not working yet. This is fixable in a single focused pass; none of the seven checks require a rebuild.
One thing that trips almost everyone up
Having the file isn't the same as it actually working
It's common for a site to have all the right files in its source code, but for the live website to still fail these checks — because of how the site is deployed. Always check the live URL (yoursite.com/robots.txt, not just the file in your project folder). If a developer tells you "it's in there," ask them to pull up the live link and show you it loads.
Next step
Run through your score. If you land under 12, you have a short, concrete list of fixes rather than a mystery. Every one of these seven items is a known, bounded fix — no guessing required.
Want a current audit or implementation plan?
Use the checklist for a self-review, then start any current audit, fix plan, scope, availability, terms, and payment request at Nymrel. JalenBuilds keeps this checklist as public educational proof.