An AI agent readiness audit is the entry-level service through which we analyze your site exactly as an AI agent would see it when trying to read it, interpret it, and act on it. We don't make assumptions and we don't rely on theory alone: we check four technical layers, from raw access to content through to how WebMCP tools actually register in a real browser. The result is a clear report, ranked by impact and effort, that you can use either to fix issues internally or as a specification for the team handling implementation. This approach lets you decide, based on concrete data, whether to fix things internally or bring in a specialized team.
What an AI agent readiness audit means
Most sites were built for human visitors and classic search engines. An AI agent has different needs: it needs unconditional access to content, tools it can call directly, structure it can query without ambiguity, and a way to discover what the site offers before crawling it in full. Our audit starts from these four needs and checks, point by point, where your site meets them and where it doesn't. The difference between being visible to people and being accessible to agents is often invisible in the visual interface, but decisive behind the technical curtain.
The four layers we check
We structure the analysis into four layers, in the order an agent encounters them:
- Access: what response code a request receives, what rules exist in robots.txt, whether relevant pages have noindex tags, and whether the firewall or anti-bot protections block legitimate agents.
- Execution: which WebMCP tools actually register in a real Chrome browser, including deferred scripts and simulated interactions, not just in static source code.
- Structure: whether headings are clear and hierarchical, whether correct structured data exists, and whether essential information can be queried without requiring visual interpretation.
- Discovery: whether a valid WebMCP manifest exists and an llms.txt file that describes what the site can offer an agent before it starts crawling pages.
What we deliver
At the end of the audit, you receive a written report, structured so it can be used directly by a technical team:
- The list of findings, ranked by impact and estimated remediation effort.
- For each finding, the identified technical cause and the proposed solution.
- The score obtained in our free validator, as a documented starting point.
- An estimate of the score after applying the recommended fixes.
You can check a first-level view of your situation at any time using our free validator, and the audit details and explains what the validator only flags.
What we cannot assess remotely
Honesty about the method's limits matters as much as the accuracy of the findings. We cannot remotely confirm behavior behind authentication, flows that depend on personalized sessions or data specific to a particular user. Nor can we guarantee how a specific commercial agent will interpret your site, because each agent has its own internal logic, which isn't public and can change without notice. What we can rigorously verify is whether the technical infrastructure is prepared to respond correctly, regardless of the agent querying it. These limits don't diminish the audit's value, they define its real boundaries, so that your expectations remain realistic.
Who this service is for
The AI agent readiness audit is useful for teams that already have a functioning site and want to understand, before investing in implementation, exactly what's missing and how much effort each fix requires. It suits companies that make decisions based on clear prioritization, not assumptions, and that prefer a documented plan before allocating internal or external resources. It's also useful for product teams that want to understand the technical impact before prioritizing a sprint dedicated to AI agent readiness.
When it's better to go straight to implementation
If your site doesn't yet have any AI agent readiness elements, if the technical structure is simple and predictable, or if your team doesn't have internal capacity for remediation, a separate audit may add an extra step without real benefit. In these situations, we recommend going straight to the WebMCP implementation service, where the initial analysis is included in the process and isn't billed separately.
How to get started
If you'd like a clear assessment of your current status before deciding on a direction, you can reach out through a quote request. We'll reply with details about the process and what the audit specifically involves for your site.