The agentic web, concretely
The agentic web is the moment when part of your visitors stop being human. A user no longer types "best waterproof running shoes" into Google to compare ten tabs. They ask an agent — ChatGPT Operator, a browser assistant, Perplexity — to do it for them. The agent opens the pages, reads the content, compares prices, and in the most advanced scenarios, adds to cart and completes the order.
This shift is massive. ChatGPT exceeds 900 million weekly active users. A growing fraction of these queries go through agents capable of browsing and acting, not just answering. Agentic commerce — purchasing triggered by an AI — is no longer a projection, it's a traffic layer being built today.
An agent is not a web user
The strategic mistake is to treat the agent as one more visitor. It isn't. A human tolerates a slow site, an overloaded design, information that has to be hunted down. The agent, on the other hand, has a functional goal: extract a piece of data, compare, act. Anything that slows down this extraction diverts it to a more readable competitor page.
An agent is not an SEO crawler
Google's crawler indexes to rank later. The AI agent reads to decide now. It has neither the time nor the patience to come back. If the data isn't immediately usable in the HTML it receives, it doesn't exist for the agent. This is Answer Engine Optimization logic taken to the extreme: not only answering, but enabling action.
Why agents don't see your site
The number one cause of agentic invisibility is simple: your content depends on JavaScript. LLMs and the majority of agents don't render JavaScript reliably. They consume the HTML returned by the server. If your application displays its prices, stock, or descriptions only after client-side hydration, the agent receives an empty shell.
This is the classic trap of poorly configured React, Vue, or Angular SPAs: gorgeous for the human, opaque for the machine. SSR or static rendering isn't a convenience optimization, it's the condition for existing in the agentic web.
AI agents are not full browsers. They parse the raw HTML returned by your server. If disabling JavaScript makes your content disappear, you are invisible to them. SSR or static rendering is non-negotiable.
The other barriers
Beyond JavaScript, several recurring blockers cut agents off from your pages:
- Missing HTML semantics:
<div>everywhere, no real<h1>,<article>,<nav>, or<button>. The agent can no longer distinguish content from decoration. - Data locked inside images: a price displayed as an image, a menu rendered in canvas. Unreadable for the machine.
- Actions hidden behind opaque JavaScript: an "Add to cart" that isn't a link or a semantic button, but an inaccessible event handler.
- Blocking AI user-agents in
robots.txtor via firewall, which shuts the door before reading even begins.
Each of these barriers turns a potential sale into an ignored page.
Making your site machine-readable
Making a site readable by agents rests on three pillars: reliable server rendering, strict HTML semantics, and explicit structured data. The goal is that every decisive piece of information — price, availability, features, reviews — is present and identifiable in the initial HTML, without interpretation.
| Element | Classic site (human) | Agentic site (machine) |
|---|---|---|
| Rendering | Client-side JavaScript | SSR or static, full HTML |
| Structure | Generic styled divs | Semantic HTML (article, nav, button) |
| Price and stock | Displayed after hydration | In the HTML + JSON-LD Schema.org |
| Navigation | Dynamic JS menus | Real, crawlable href links |
| Actions | Opaque JS handlers | Semantic forms and buttons |
The decisive role of Schema.org
Schema.org structured data is the common language between your site and agents. A Product markup exposes the price, currency, availability, and reviews without ambiguity. A FAQPage markup provides directly usable question-answer pairs — and it's also a strong signal for AI Overviews, which now trigger on more than half of Google queries.
When an agent has to compare ten products in a few seconds, it systematically favors the pages where this information is declared in JSON-LD over those where it has to be guessed from the text. Markup isn't a bonus: it's what tips the agent's choice toward you.
Semantics as a readability contract
A <button> is an action, a <nav> is navigation, an <article> is main content. These tags aren't cosmetic: they form the contract that lets the agent map your page in a single read. Replacing decorative <div> elements with real semantic tags is often the fastest win.
Making your site actionable
Being readable isn't enough in the agentic web: you have to be actionable. The difference between a site that gets cited and a site that converts with agents comes down to one thing — can the agent complete the action without human intervention? Adding to cart, booking a slot, submitting a quote form: these journeys must rely on standard, predictable elements.
Enable SSR or static generation. Verify that prices, stock, and content appear in the initial source, without depending on the client.
Declare your Product, Offer, FAQPage, and Organization types. Fill in price, currency, availability, and reviews. This is the agents' direct access to your data.
Your action buttons are real <button> or <a href> elements, your forms real <form> with explicit labels. Banish opaque JavaScript handlers.
Allow the relevant AI crawlers in your robots.txt. Expose an llms.txt that points to your key pages and your structured data.
Avoid random URL changes, aggressive CAPTCHAs, and superfluous steps. A predictable journey is a journey an agent can complete.
Off-site authority matters too
An agent doesn't decide based on your page alone. It relies on what the web says about you. Ahrefs' analysis of 200,000 domains (December 2025) shows that off-site brand mentions correlate far more with AI citations than Domain Rating: YouTube leads (0.737), followed by Reddit and Wikipedia, versus only 0.266 for DR. Wikipedia alone accounts for 47.9% of ChatGPT citations.
Off-site presence — Wikipedia, Reddit, YouTube — weighs more than your domain authority for being chosen by an agent. Optimize the page and the brand ecosystem around it.
Concretely: an agent hesitating between two equivalent suppliers will decide in favor of the one whose reputation is attested elsewhere. Technical readability opens the door; off-site authority makes the final choice.
Measuring your agentic readability
Measurement starts with a free, immediate test: disable JavaScript in your browser and reload your key pages. If the content, prices, or buttons disappear, agents ignore them too. It's the most brutal and revealing diagnostic.
The verification protocol
Go further with four checks:
- Raw HTML: query your pages with
curlor Google's URL inspection tool. Compare what the server receives with what the human sees. The gap is your agentic blind spot. - Schema.org validity: run your pages through Google's Rich Results Test. Every markup error is a piece of data the agent won't be able to use.
- Action accessibility: verify that your purchase or contact journeys rely on semantic elements navigable without a mouse.
- Server logs: spot AI user-agents (GPTBot, PerplexityBot, ClaudeBot) in your logs. Their presence — or absence — reveals whether you're already on their radar.
To structure this approach end to end, our France GEO Guide details the technical checklists and implementation priorities. And if you want to understand the underlying logic behind all of this, AEO lays its foundations: answer directly, structure, prove.
What measurement reveals
The agentic web separates two categories of sites. Those that expose their data cleanly — full HTML, Schema.org, semantic actions — capture a rapidly growing layer of software traffic. Those that rely on JavaScript and unpredictable journeys become invisible to millions of users who now delegate their browsing. The good news: most of the fixes are technical, measurable, and achievable. The bad news: your most advanced competitors are already on it.
Free GEO audit: we test your server rendering, your Schema.org, and your agentic readability, then hand you a prioritized action plan.
Questions fréquentes
What is an AI agent for the web?+
An AI agent is autonomous software capable of browsing the web, reading pages, filling out forms, and triggering actions on behalf of a user. ChatGPT Operator, Perplexity, and browser shopping assistants are examples. Unlike a traditional SEO crawler, the agent acts: it compares, selects, and sometimes buys.
Do AI agents execute my site's JavaScript?+
Most AI agents and crawlers don't render JavaScript reliably. They read the HTML returned by the server. If your content, prices, or buttons only appear client-side after hydration, the agent sees a blank page. Server-side rendering (SSR) or static rendering is therefore essential to be readable.
Is Schema.org really useful for AI agents?+
Yes. Schema.org structured data gives agents direct, unambiguous access to your prices, availability, reviews, and possible actions. An agent that has to compare ten products favors pages where this information is explicit in JSON-LD rather than buried in layout text.
How do I know if my site is readable by AI agents?+
Disable JavaScript in your browser and reload the page: if the content disappears, agents don't see it either. Also inspect the raw HTML via 'curl' or Google's URL inspection tool. Verify that prices, stock, descriptions, and action links are present in the initial source.
