Free guide · 5-minute read

The 12-Point AI Search Readiness Checklist.

The same items I check first when I open a new audit. None of these will get you cited — but skipping them will get you ignored. Each one takes about 15 minutes.

You can DIY the whole list. The harder part — reading what AI actually says about your business, hand-verifying every quote, and mapping your real competitor set — is where the audit comes in.

By Owen Kurth, founder · BeCited · Save this page (⌘P → Save as PDF) for your team.
Tier 01 · Technical

Make AI find you.

If retrieval bots can’t reach your pages, none of the rest matters. Four foundations.

01
Allow AI retrieval bots in robots.txt.

Most sites block crawlers without realizing they’re also blocking the bots that fetch live answers for ChatGPT, Perplexity, and Claude. If those bots can’t crawl, your content can’t be cited — even if a model technically knows your brand.

Do thisOpen yourdomain.com/robots.txt. Allow at minimum: OAI-SearchBot, ChatGPT-User, PerplexityBot, Claude-SearchBot, Google-Extended. If you’re unsure what to block, allow all of them — you can tighten later.
02
Publish an llms.txt at /llms.txt.

A short, structured map of what your site offers and where the canonical pages live. Treat it as a small README for AI engines — not a spec, just a signal that you’re ready to be read.

Do thisAdd a plain-text file at yourdomain.com/llms.txt. Lead with: business name, a one-line description, and 5–10 bullets pointing to your most important pages with one-line descriptions. Bonus: add an llms-full.txt with the full text of those pages.
03
Keep a current sitemap.xml.

A sitemap tells crawlers what exists, when it changed, and which pages matter. AI engines use the same signals search engines use — if your sitemap is missing or stale, fresh content gets discovered slower.

Do thisConfirm yourdomain.com/sitemap.xml returns valid XML and lists every published page. Most CMSs (WordPress, Webflow, Shopify, Squarespace) generate this for you — just verify it’s referenced in your robots.txt with a Sitemap: line.
04
Add Organization JSON-LD to your homepage.

Structured data is the most reliable way to tell an AI engine your name, what you do, where you operate, and how you connect to your other profiles — in a format the engine doesn’t have to guess at.

Do thisDrop a JSON-LD Organization (or LocalBusiness if you serve a region) script in the <head>. Required fields: name, url, logo, sameAs (an array of your social, review, and Wikipedia profiles), description. Validate at validator.schema.org.
Tier 02 · Content

Make AI quote you.

Discoverable isn’t enough. AI answers favor pages it can lift a clean sentence out of.

05
Lead every page with the answer.

Engines pull the first 1–2 sentences of a relevant chunk and quote it. If your opening paragraph is a marketing windup, you’re handing the citation slot to someone with a tighter intro.

Do thisRead the first sentence of each top page out loud. Does it answer the question the page is trying to rank for? If not, rewrite it as a direct, specific statement. The brand monologue can come second.
06
Use semantic HTML, not div soup.

Engines extract content from semantic tags (main, article, section, h1–h3, ul, ol) more reliably than from generic div stacks. It’s the difference between handing them a labeled folder and a pile of paper.

Do thisOpen one important page’s source. Find the main content. Is it inside <main> and <article>? Are the headings real <h1>/<h2> tags or styled divs? If everything is div class="heading", it’s costing you.
07
Add a real FAQ section.

FAQs are the format AI engines understand best — question header, direct answer, repeat. This is the single highest-leverage content shape for AI visibility.

Do thisPick the 5 questions buyers actually ask before they hire you. Write 1–3 sentence direct answers. Mark them up in FAQPage JSON-LD or use plain <details>/<summary> tags — both work.
08
Pack in numbers, dates, and dollars.

AI answers love specifics. “We’ve served Seattle homeowners since 2014” quotes cleaner than “trusted local provider for years.” Vague claims rarely make it into citations; specifics often do.

Do thisAudit your top 3 pages for any line that says “many,” “leading,” “trusted,” or “years of experience.” Replace each with the actual number. If you don’t have the number, find it once and write it down forever.
Tier 03 · Reputation

Make AI trust you.

Most AI citations come from third-party sources, not your site. Get on the right ones — and look the same on each.

09
Claim your top 3 review-platform profiles.

AI engines lean heavily on review aggregators when answering category questions. An unclaimed Yelp or G2 profile is a free win for your competitors — the AI defaults to whoever’s actively present.

Do thisFor local services: Google Business Profile, Yelp, BBB. For SaaS: G2, Capterra, TrustRadius. For consumer products: Amazon, Trustpilot, the relevant niche publications. Claim, fill out completely, add photos, ask 3 happy customers for a review.
10
Stop using your name three different ways.

If you’re “Acme Co.” on your homepage, “Acme Cleaning” on Yelp, and “Acme Inc.” on your invoices, AI engines will sometimes count those as three different businesses — and split your visibility across all three.

Do thisPick one canonical name. Use it identically on: your website’s schema, Google Business Profile, every review platform, your social handles, your email signature. Same wording, same punctuation, every time.
11
Get into one third-party listicle.

A single mention in a “Best [your category] in [your city/segment]” post by a respected source moves the needle more than a month of on-site SEO. AI engines are pattern-matchers — if three trusted lists mention you, you’re a recommendation.

Do thisSearch “best [your category] [your geography or buyer]” on Google. Find the 5 articles ranking on page one. Email the authors with a one-paragraph case (specific differentiator, evidence). Don’t pay; offer a real story.
12
Add an “About” or team page with credentials.

AI engines weight expertise signals heavily — named people, real credentials, years in the field. A faceless About page reads as untrustworthy. A specific one (“Owen Kurth, 8 years in service operations”) reads as legitimate.

Do thisAdd real names, photos, and 1-line credentials to your About page. Use Person JSON-LD with jobTitle, worksFor, and sameAs linking to LinkedIn. If you have certifications or licenses, name them on the page.
When you’re done

You’ve handled the easy 80%. The other 20% is what BeCited does.

What this checklist can’t do: read what AI actually says about your business across all four engines, hand-verify every quote, separate real competitors from national chains, and trace each gap back to its root cause. That’s the BeCited Full Audit — 100–300 buying-intent prompts, four engines, one named analyst, one week.

© 2026 BeCited · Saambaa LLC (Washington) becited.io · Terms · Privacy