This AI search optimization checklist covers the twenty items that decide whether AI assistants and AI search engines - ChatGPT, Claude, Gemini, Grok and Perplexity among them - can find, understand, verify and recommend a B2B company. Work through the four groups in order - accessibility first, then clarity, then evidence, then citations - because each group depends on the one before it.
Technical accessibility (items 1-6)
AI systems can only recommend what they can read. See also how AI reads your website.
- Key content server-rendered, not JavaScript-only.
- AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended) allowed in robots.txt - blocking them removes you from AI search.
- Organization schema with name, category, logo and sameAs links.
- llms.txt published, listing your most important pages.
- Clean HTML headings that match visible text.
- Critical claims in text, never only in images or PDFs.
Entity clarity (items 7-11)
Models recommend companies they can classify. Start with a category sentence a machine can parse.
- One category sentence on the homepage a machine can parse.
- The same sentence on About, LinkedIn and every directory profile.
- Consistent legal and brand naming everywhere.
- Founders and key people named and linked.
- No contradictions between your site and external profiles - inconsistency reads as unreliability.
Buyer-intent content and evidence (items 12-16)
AI answers are built from the questions buyers ask.
- Pages answering comparison questions ("X vs Y") in your category.
- Use-case pages phrased the way buyers phrase them.
- Named customers and concrete numbers instead of anonymous superlatives.
- Case studies with dates and verifiable outcomes.
- Public methodology or process page - the single most citable asset a service company can own.
EntityRise publishes its own measurement methodology for the same reason.
Run the $49 audit to see what AI says about your company today and which gaps to fix first →
External citations (items 17-20)
Live AI answers are composed from cited sources. See AI citations explained.
- Accurate presence on the review and comparison sites AI engines cite for your category.
- Industry directory listings with consistent data.
- Mentions in third-party articles buyers' questions surface.
- Monitoring routine that re-checks the same buyer questions on a schedule, for example by repeating the audit, because answers drift as models and sources change.
Frequently asked questions
How many items do most B2B sites fail?
In audit practice, most companies fail a large share of the accessibility and evidence groups - not because the fixes are hard, but because nobody had measured, so nobody owned the problem.
Which items matter most?
Sequence beats cherry-picking: accessibility failures silently cancel everything else, so fix 1-6 first, then clarity, then evidence, then citations.
Is this the same as an SEO checklist?
They overlap on technical hygiene, but items like llms.txt, AI-crawler access, category classifiability and citation mapping are specific to AI visibility. The differences are explained in our GEO vs AEO guide.
Kamil Kępiński · Founder of EntityRise and Smophy Labs Inc.
Kamil Kępiński is the founder of EntityRise and Smophy Labs Inc. He writes about how models choose the companies they recommend and how to measure a company's visibility in AI answers.
About the authorXLinkedInThe methodologycontact@entityrise.ai
Twenty items is a quarter of side-of-desk work, so start with the ones the evidence points to.
EntityRise measures what AI assistants say about your company and ranks the findings by priority, so your own team or agency knows what to fix first.
Run the audit · $49 · For agencies: packs from $28 per audit with white-label reports