Agent-Ready Website Checklist: 15 Points for 2026
Score your website against these 15 points — from semantic HTML and schema markup to llms.txt and a booking API — to find out if AI agents can find, understand, and recommend your business.
Agent-Ready Website Checklist: 15 Points for 2026
AI agents are becoming the new visitors to your website. Not in five years. Right now.
ChatGPT Browse, Google AI Overviews, Perplexity, and soon dozens of other AI tools crawl the web to provide answers, recommend services, and even book appointments. The question isn't whether they'll visit your website. The question is: do they understand what they see?
Most SMB websites are built for humans. That's fine — but no longer enough. Your website also needs to be readable by machines. We call that Agent-Ready.
Here's the checklist. 15 points, grouped by priority. Score yourself honestly.
Must-Have (without these you're invisible)
1. Semantic HTML
What: Use the correct HTML elements for the right content. <header> for your header, <nav> for navigation, <main> for main content, <article> for articles, <section> for sections.
Why: AI models use HTML structure to understand what's important on your page. A <div> says nothing. An <article> says everything.
Check: View your source code (Ctrl+U). Do you see mostly <div> tags? There's work to do.
2. Schema.org markup
What: Structured data in JSON-LD format that tells search engines and AI what your business does.
Why: Schema markup is the standard language between your website and AI systems. Without it, they have to guess. With it, they know for certain.
Check: Paste your URL into the Google Rich Results Test. No results? No schema markup.
3. Fast loading time (< 2 seconds)
What: Your website loads completely within 2 seconds, including on mobile.
Why: AI crawlers have timeouts. Slow websites get skipped or incompletely crawled. Google's Core Web Vitals are also a ranking factor.
Check: Test on PageSpeed Insights (pagespeed.web.dev). Score below 80? Action needed.
4. Mobile-first design
What: Your website works perfectly on smartphones. Not "it fits," but "it works excellently."
Why: Google has indexed mobile-first since 2021. More than 60% of all website visits come from mobile (Statista, 2025).
Check: Open your website on your phone. Do you need to zoom or scroll horizontally? Not mobile-first.
5. robots.txt that doesn't block AI crawlers
What: Your robots.txt file doesn't block AI crawlers (GPTBot, Google-Extended, PerplexityBot, ClaudeBot).
Why: Many default robots.txt configurations block all bots except Googlebot. That means AI models can't read your website.
Check: Go to yoursite.com/robots.txt. Does it say "Disallow: /" for unknown user-agents? You might be blocking AI crawlers.
Should-Have (this is how you differentiate)
6. llms.txt
What: A file in your root directory (yoursite.com/llms.txt) that gives AI models a summary of your business, services, and how they can recommend you.
Why: Think of it as robots.txt, but specifically for AI. It's a direct briefing to every AI that visits your website.
Check: Go to yoursite.com/llms.txt. 404? You don't have one.
7. Open Graph tags
What: Meta tags that determine how your page looks when shared on social media or cited by AI.
Why: AI models use Open Graph data (og:title, og:description, og:image) to summarize and present your page.
Check: Paste your URL into the Facebook Sharing Debugger or opengraph.xyz. No preview? No Open Graph.
8. FAQ schema
What: Structured FAQ data that Google displays as expandable questions in search results.
Why: FAQ schema is doubly valuable: it delivers rich snippets in Google AND it gives AI models direct answers to cite.
Check: Google Rich Results Test. Look for "FAQ" in the results.
9. Clear service descriptions
What: Every service you offer is clearly described on its own page or section, with name, description, target audience, and preferably price.
Why: AI agents match search queries to services. "Who offers accounting in Utrecht?" — if your service isn't clearly described, you get skipped.
Check: Can someone visiting your website for the first time understand what you offer within 10 seconds? Can a machine?
10. Contact information directly accessible
What: Name, address, phone number, and email address appear on every page (footer) and on a dedicated contact page.
Why: AI agents that make recommendations also want to provide contact details. If those are hidden behind a form, you get skipped for a competitor who shows them directly.
Check: Is your phone number in the footer? Is your address findable without clicking?
Nice-to-Have (this is how you lead)
11. Author credentials
What: Content (blogs, articles, guides) includes the author's name with a brief bio, title, and preferably a link to a LinkedIn profile.
Why: Google's E-E-A-T framework (Experience, Expertise, Authoritativeness, Trustworthiness) weighs author credentials heavily. AI models do the same.
Check: Is there an author name on your blog posts? Does that person have a visible profile?
12. Review signals
What: Reviews and ratings are visible on your website, preferably with AggregateRating schema markup.
Why: Social proof. AI models cite businesses with demonstrable customer satisfaction more readily.
Check: Are there reviews on your website? Do they have schema markup?
13. Structured content
What: Your content is built with clear headings (H1, H2, H3), paragraphs, lists, and tables. Not a wall of text.
Why: AI models extract information based on heading structure. An H2 "Our Pricing" followed by a table is immediately usable. A paragraph with prices buried in it is not.
Check: Does every page have exactly one H1? Do H2s and H3s follow a logical hierarchy?
14. API readiness
What: Your website has (or can have) an API endpoint that allows AI agents to directly perform actions — book an appointment, check availability, request a quote.
Why: The next step after "being found" is "being directly served." AI agents that can book an appointment without a form convert up to 3x better.
Check: Can a machine currently do anything on your website besides reading? This is forward-thinking — but it's coming.
15. WebMCP preparation
What: Model Context Protocol (MCP) is an emerging standard that allows AI models to communicate directly with your website — not through scraping, but via a structured interface.
Why: MCP is to AI agents what APIs were to apps: the standard way AI agents communicate with services. Being prepared early gives you a head start.
Check: This is still emerging technology. Awareness is the first step.
How do you score?
Add up your points:
| Score | Status | Action |
|---|---|---|
| 13-15 | Agent-Ready | You're ahead of the curve. Maintain and monitor. |
| 10-12 | Almost there | A few targeted improvements and you're set. |
| 7-9 | Work to do | The basics are there, but you're missing crucial elements. |
| 4-6 | Falling behind | Your competitors are ahead. Time for action. |
| 0-3 | Invisible | AI models don't know you. Start today. |
What does it cost to become Agent-Ready?
The must-haves (points 1-5) are largely free to implement. They mainly require knowledge and time. The should-haves (points 6-10) require a bit more technical implementation. The nice-to-haves (points 11-15) are where you really make a difference — and where professional help makes the difference between "we have it" and "it works."
Conclusion
The world is changing. Websites built only for humans will soon no longer be found — not by Google, not by ChatGPT, not by any AI.
Being Agent-Ready isn't a luxury. It's the next standard. And the businesses that start now will have an advantage that's hard to catch up to.
Scoring below 10? Request a free AI Readiness Scan. We'll walk through all 15 points for your website and give you a concrete action plan. No obligations, no sales pitch. Just clarity on where you stand.
Klaar om te automatiseren?
Plan een vrijblijvend gesprek en ontdek wat AI-automatisering voor jouw bedrijf kan betekenen. Of doe eerst de 2-minuten AI Readiness Scan.
Plan een gesprek