Kaliber Group — Performance Marketing, Augmented by AI

How to Fix AI Readability on Webflow (2026)

19 May 2026

Why this matters for Webflow sites

The same Designer that surfaces typography and spacing doesn't tell you whether your robots.txt allows ClaudeBot, whether Organization schema has a logo, or whether your hero element is set to H1 or H2. AI engines read all three. Owners rarely check.

A second pattern is agency-built sites handed over with the SEO panels skipped or filled with placeholder text. The Designer lets you ship a beautiful site without ever opening the SEO tab. Sites in this state typically score 25-40 points below their design quality would suggest.

The fixes are ordered by how much each moves the audit score, not by where each setting lives in the Designer.

What Google said in May 2026

Google published an updated AI Search guide on May 7, 2026. The headline: optimising for generative AI search is optimising for the search experience, and thus still SEO.

Three claims worth knowing:

• No special files needed. llms.txt and agents.md give no benefit for Google's AI surfaces. They may still help niche LLM tools that read them, but the broad-adoption case is dead.
• No content chunking needed. Markdown mirrors, sitemap.md, and similar parallel formats are unnecessary. Google's systems handle multi-topic pages natively.
• No special schema needed for AI. Schema still helps traditional rich results and is still consumed by ChatGPT, Claude, and Perplexity. But adding schema because of AI search is no longer the angle.

Google also stopped rendering FAQ rich results the same day. The markup still parses; the SERP feature is gone.

1. Allow AI crawlers in robots.txt

The single highest-impact fix. AI engines can't read what they can't fetch. Webflow's default robots.txt has no explicit allow rules for AI crawlers — and depending on how the site was configured, an unscoped Disallow may have crept in.

The robots.txt editor lives in Project Settings on all paid Site plans. Free webflow.io subdomain sites can't edit it.

The block to add

Paste this into the Robots.txt field. Webflow appends the Sitemap line automatically, so you don't need to add it.

User-agent: GPTBot
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: anthropic-ai
Allow: /

User-agent: CCBot
Allow: /

User-agent: Google-Extended
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Applebot-Extended
Allow: /

User-agent: cohere-ai
Allow: /

User-agent: Bytespider
Allow: /

To allow search indexing but block AI training: set Disallow: / on Google-Extended and Applebot-Extended only. Those control training data without affecting search crawlers.

2. Fix heading hierarchy

AI engines use headings to map the page. The audit flags pages with more than one H1, skipped levels (H1 jumping to H3), or fewer than five H2/H3 headings on a substantive page.

The Webflow-specific bug: Symbols. A Hero Symbol may contain an H1, then someone adds another Heading element set to H1 on the page. Two H1s, not obvious from the visual layout.

The fix

One H1 per page. H2 for body sections, H3 for subsections, H4 only when needed. Don't skip levels.

Open the Navigator panel (left sidebar). Expand every Symbol. Click each Heading element and check the Tag dropdown in the Element Settings panel. If the page is built from a Webflow Template Marketplace theme, audit it after launch — template defaults often set wrong heading levels and most designers never look.

3. Write meta descriptions that say something

AI engines read meta descriptions to decide whether to surface a page in an answer summary. Two patterns fail on Webflow: descriptions under 100 characters, and template defaults nobody rewrote.

A description that passes names at least one of: a specific category, a named differentiator, or a location. "Award-winning aesthetic treatments in Singapore" passes (category + geography). "A leading clinic dedicated to your beauty journey" fails — nothing concrete to extract.

4. Add Organization schema

Organization schema is how AI engines link your domain to your brand. Without it, they often cite a competitor instead of you when answering questions in your category.

Webflow has no built-in generator. Add it via Custom Code so it renders site-wide.

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Your Business Name",
  "url": "https://yourdomain.com",
  "logo": "https://yourdomain.com/path-to-logo.png",
  "sameAs": [
    "https://linkedin.com/company/your-business",
    "https://twitter.com/your-business"
  ],
  "contactPoint": {
    "@type": "ContactPoint",
    "telephone": "+65-6000-0000",
    "contactType": "Customer Service"
  }
}
</script>

If your business has a physical location, swap "@type": "Organization" for "@type": "LocalBusiness" and add an address block. AI engines treat LocalBusiness as a stronger signal for geo-relevant queries.

5. Add Service or Product schema

Organization schema says you exist. Service or Product schema says what you do. Every major service page should have its own Service block. Every product page should have Product with at minimum name, description, image.

Webflow Ecommerce auto-renders basic Product schema on product templates. Service schema almost always needs adding by hand.

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Service",
  "name": "Performance Marketing",
  "provider": {
    "@type": "Organization",
    "name": "Your Business Name"
  },
  "areaServed": "Singapore",
  "description": "Senior-led performance marketing for B2B and consumer brands.",
  "serviceType": "Marketing Services"
}
</script>

6. FAQ schema after Google's May 2026 change

Short answer: keep using it on pages that are genuinely Q&A. Stop using it as a SERP-snippet hack on pages that aren't.

The SERP rich result is gone. The markup still parses cleanly. Non-Google LLMs (ChatGPT, Claude, Perplexity) appear to still consume FAQ schema when summarising — Google hasn't confirmed whether their own AI surfaces use it.

Webflow has no native FAQ widget that emits FAQPage schema. Custom Code is the only path.

7. Canonical edge cases

Webflow sets canonicals on native pages automatically — fine for most sites. Two cases need a manual override: content syndicated from elsewhere (canonical should point to the original), and paginated CMS Collection list pages where Webflow's default canonical may not match your URL strategy.

8. Make your content concrete enough to be cited

This is the check Google would say matters most. The audit scores three things on the actual content: specificity (does the page name real products, services, or methods?), differentiation (does it distinguish you from alternatives?), and credibility (does it name real clients, results, proof?).

Two Webflow-specific patterns to fix:

Hero copy borrowed from the template. Many sites launch with template hero text ("Beautiful design, made simple") nobody rewrote. AI engines have nothing to extract. Replace with a specific claim: what you do, for whom, with what outcome. "Senior-led performance marketing for SEA consumer brands. Most engagements move CAC down 30%+ in 12 weeks" gives an AI engine four citation hooks. The template version gave zero.

Image-heavy pages with thin text. Webflow's design control encourages visual storytelling — good for users, bad for AI parsing. AI engines extract from text, not images. If your service page is 90% imagery with three short captions, you're not citable. Add 200-400 words of body copy per major service. Use H2/H3 to structure it. Webflow's Rich Text element renders cleanly.

9. Verify everything is actually working

Three free tools. Take five minutes per page.

Schema: paste your page URL into Google's Rich Results Test. Critical for Webflow — Custom Code only renders on the PUBLISHED site, not the Designer preview. Always test the live URL after Publish. validator.schema.org is stricter — use it when the Google tool says everything's fine but something still looks off.

Robots.txt: visit yourdomain.com/robots.txt in incognito. You should see your block plus the auto-appended Sitemap line. If you only see Webflow's default User-agent: *, the changes saved but you never republished.

End-to-end: re-run the AI Readability Checker. Each fix should move the score visibly.

What to fix first — a half-day order of operations

First (10 minutes). Update robots.txt to explicitly allow AI crawlers. Biggest score lift per minute spent.

Second (20 minutes). Add Organization schema site-wide via Project Settings → Custom Code → Head Code. Publish. Confirm via Rich Results Test.

Third (an hour). Rewrite homepage hero and top three service pages for concrete claims. Vague taglines out. Specific category, audience, outcome in.

Fourth (an hour). Fix heading hierarchy on top pages. Navigator panel → expand Symbols → one H1 per page.

Fifth (an hour). Rewrite meta descriptions on top pages — 120-160 characters, one concrete claim each. For CMS Collection pages, bind the field and fill it per item.

The other fixes (Service/Product schema, FAQPage, canonical tweaks) are useful but smaller. The first five account for most of the score lift.

One pass, lasting effect

None of these fixes need ongoing maintenance once set up. The robots.txt block sits in Project Settings forever. Schema renders on every page via Custom Code. Heading hierarchy stays fixed unless someone explicitly reverts it. Meta descriptions, once written well, persist.

That's why it's worth getting right on the first pass. Sites that scored poorly in March usually still score poorly in October — nothing changed.

If your audit surfaces fixes that need more than the Designer can give you — a custom interaction that interferes with structured data, a CMS template that won't accept dynamic schema, content that needs proper rewriting rather than tweaking — that's the work we take on. Get in touch, or read about the GEO & AI Search service.

Frequently asked questions