Agent Readiness: Is Your Website Ready for AI Shoppers
For twenty five years a website was built for one visitor: a person with a screen, a mouse or a thumb, and the patience to look around. A second visitor has arrived. AI agents now visit websites on behalf of people, to find facts, compare options and complete tasks. They do not look around. They read what they can parse, call what they can reach, and leave when they cannot. Agent readiness is the question of whether your site works for this visitor. Most do not, for reasons that are easy to list and increasingly important to fix.
How an agent visits
An agent arrives with a goal: find a hotel near the old town with a pool, confirm it allows children, check a date, book it. It fetches pages and extracts facts. It looks for structured data, because that is the fastest way to get facts without guessing. It looks for actions it can take: an availability check, a booking endpoint, a form it can understand. If it finds what it needs it proceeds. If a page is a blank shell that only a browser can render, a brochure made of images, or a booking flow designed for a human to click through, it either gives up or moves to a competitor it can use.
The readiness checklist
Can the content be read without a browser. View the source of your key pages. If the facts are present as text, an agent can read them. If the page is empty until scripts run, it may not be.
Are the facts stated, not implied. Hours, prices, policies, amenities, location, contact details should exist as plain statements, not atmosphere. An agent cannot act on "open until late".
Is there structured data. Schema.org markup for the business, its outlets, hours, policies and FAQ entries hands the agent clean facts and declares your identity. It is the single highest return item on the list.
Is every fact consistent. The agent may check your site against a directory and a booking platform. Disagreement makes it hesitate. One version of every fact, everywhere.
Is there a machine readable map. A sitemap for crawlers and an llms.txt file for models help an agent find the pages that matter without wandering.
Can an agent act. The step beyond reading. A booking engine with a documented interface, or an MCP server exposing availability and reservation tools, lets an agent finish the task with you rather than handing the person a link.
Are the images and documents readable. Menus as photographs, brochures as flipbooks and prices in downloadable images are invisible. PDF menus are read poorly or not at all, and a column layout often scrambles which price belongs to which item. Provide the same information as a page in the formats machines read: structured markup, question and answer pairs, schema.org and a knowledge graph.
Is the identity clear. A canonical page, consistent naming, and links to your profiles elsewhere let the agent confirm it has found the right business.
What changes when you are ready
An agent ready site is chosen more often, because the agent can verify you match the request. It is described more accurately, because the facts came from you rather than from inference. And it converts, because the agent can complete the booking instead of abandoning the task. The businesses that benefit most are not the largest. They are the ones whose facts are clearest and whose actions are reachable.
What does not matter to an agent
Visual design, animation, hero imagery and brand voice are for people, and they still matter for people. An agent ignores all of it. This is good news: readiness does not require redesigning your site. It requires adding a machine readable layer underneath the one people see, and making sure the facts in that layer are complete, consistent and current.
Where to start
Pick the ten questions a guest most often asks. Make sure each has a plain text answer on your site, marked up with structured data, that agrees with every listing. Check the pages render without scripts. Add a sitemap and an llms.txt file. Then look at whether an agent could book, and what it would take to make that possible. That sequence takes a site from invisible to operable.
The takeaway
AI agents are a new kind of visitor that reads facts and takes actions rather than browsing. A site is agent ready when its facts are readable, structured, consistent and reachable, and its key actions can be called by a machine. The design stays for people. The readiness is a layer beneath it, and it is quickly becoming the layer that decides who gets chosen.