Evaluating "agent-ready websites" with isitagentready.com — New standards for corporate site redesign client projects 2026

From "searching on Google" to "asking AI agents"—the starting point for gathering information is shifting. As of 2026, use cases where agents such as ChatGPT, Claude, and Gemini browse corporate websites on behalf of users, extract information, fill out forms, and complete bookings are expanding rapidly.

In the midst of this trend, isitagentready.com, released by Cloudflare, is a free tool that diagnoses whether your website is designed to be "read" by AI agents. This article breaks down its diagnostic criteria and outlines the new design standards to adopt in corporate website client renewals.

The difference between "human-friendly" and "agent-readable"

Over the past decade, corporate websites have been optimized to be "visually appealing to humans." SPAs, CSS animations, and hero images all enhance the human visual experience, but they become barriers when viewed by AI agents.

As we wrote in Structured data implementation guide in the era of AI search, structuring data has shifted from "nice to have" to "disadvantageous without it." In the agent era, websites are further questioned on "whether an agent can understand and operate all site functionality."

Evaluation axes of isitagentready.com

Here we summarize the six primary axes evaluated by the diagnostic tool.

A. Machine-readable content

• Whether core content is served via SSR / SSG
• Whether there is a <noscript> fallback
• Whether semantic HTML (<article>, <section>, etc.) is used

Content rendered only via JS is equivalent to an "empty page" from an agent's perspective.

B. Structured data

• Schema.org markup such as Organization / Product / Service / FAQ
• Structured @id hierarchy in JSON-LD
• Breadcrumbs / Sitelinks markup

C. agents.txt / llms.txt

It checks for the presence of agents.txt / llms.txt, which can be considered the AI agent equivalent of robots.txt. This is a new standard for declaring "which pages agents should prioritize reading" and "which features they are allowed to use."

D. AI-friendly forms

• Whether forms are placed via SSR and have <label> and aria-* configured properly
• Whether name attributes are machine-readable (autocomplete values such as given-name and email)
• Whether the form's purpose is clearly specified with <form aria-label>

E. Rate limiting and agent identification

• Whether there is a mechanism to identify AI agent UAs
• Whether it can block excessive traffic while returning appropriate responses to legitimate agents

F. Performance

• Whether TTFB (Time to First Byte) is fast
• Whether core content is rendered before LCP

The performance standards outlined in Designing corporate websites with Web Platform Baseline 2026 remain just as vital in the agent era.

Top 5 items to fix first in client renewals

Based on isitagentready findings, here are the five items we always address in our renewal proposals.

1. Migration to SSR / SSG

If a CSR SPA is currently deployed, switching to SSR / SSG is the top priority. Options include Astro, Next.js (App Router), and TanStack Start. This simultaneously resolves agent compatibility and improves Core Web Vitals.

2. Systematizing Schema.org

Standardize @id across the entire site,

• Company info → Organization
• Services → Service
• Locations → Place
• Careers → JobPosting
• News → Article

and configure them in JSON-LD.

3. Deploying agents.txt / llms.txt

Examples:

# agents.txt
User-Agent: *
Allow: /
Sitemap: https://example.com/sitemap.xml
Contact-Form: https://example.com/contact
Pricing: https://example.com/pricing

User-Agent: GPTBot
Crawl-Delay: 1

Explicitly declare "pages you want agents to visit" and "admin consoles you do not want them to touch."

4. Headless contact forms

Decouple form submissions as an independent API (POST /api/contact) to support both HTML forms and AI agent submissions.

<form
  action="/api/contact"
  method="post"
  aria-label="お問い合わせフォーム"
>
  <label for="email">メールアドレス</label>
  <input id="email" type="email" name="email" autocomplete="email" required>
  <!-- ... -->
</form>

Exposing it as an API enables agents to complete inquiries on behalf of users.

5. Consolidating company info into a single canonical page

Normalize the company overview, leadership message, case studies, and contact details into a single section, and reference it from other pages using <link rel="canonical">. This creates a structure where agents can instantly understand who the company is and what it does.

Conclusion — The renewal benchmark shifts from "human" to "human + AI"

Until now, corporate site renewals were judged by "how they look to humans." From 2026 onward, "how readable they are to AI agents" becomes a parallel benchmark.

Note that where to begin among the five items above varies significantly from site to site. If the site is a CSR SPA, migrating to SSR / SSG is the bottleneck; if it is already statically generated, simply structuring the JSON-LD @id schema and adding agents.txt can change the diagnostic results. Start by diagnosing your site on isitagentready.com, see which areas fall short, and then determine your renewal scope.

If you find it difficult to prioritize based on the diagnostic results or want to determine what is practical given your current website architecture, please reach out via our contact form with your diagnostic URL and existing technical stack.