AI Website audits are now available within Google Chrome DevTools. Having a website that is accessible to AI tools such as Google Gemini, ChatGPT, Co-Pilot, etc. are increasingly important. However, most marketers and business owners don’t have a way to check if their site is accessible nor know where to start to make it accessible.
Thankfully, Google has been making a lot of updates on their analytics products of late to give insight into AI performance, and now Google Chrome DevTools added a feature to score how accessible a website is to AI.
Google Chrome DevTools
Google Chrome DevTools is a suite of tools built directly in the Chrome Browser that lets developers and marketers inspect and optimize websites. One of the tools is Lighthouse which historically featured reports on performance, accessibility, best practices and SEO. Now it added a new report for Agentic Browsing.
What’s included in the Agentic AI Report?
The Lighthouse agentic browsing report has three core checks to ensure sites are accessible for AI tools, which include: LLMS.TXT, Cumulative Layout Shift and Accessibility Tree.
LLMS.TXT
Having an optimized LMS.TXT file is very important to ensure accessibility.
As listed on our LLMS.TXT blog post, LLMS.txt is like a map, to help guide AI crawlers to specific content you want to exemplify. Thanks to its ability to process language, you can feed AI business success stories, achievements, and services. When a chatbot receives a query from a user, it scours the internet for information, if it sees that your website is optimized for efficient viewing with LLMS.txt, it reviews the content you highlighted, and reports it back to the user.
Cumulative Layout Shift
Cumulative Layout Shift measures the movement of visible elements. If there is a high score, it means the visual elements of the site are unstable, which is poor for the user experience, but can also hurt the clarity of the website for both AI crawlers and search engines. In this case, you want a low Cumulative Layout Shift score.
Accessibility Tree
As quoted within the Chrome for Developers blog:
“Agents rely on the accessibility tree as their primary data model. Lighthouse filters a specific subset of accessibility audits that are critical for machine interaction
This includes:
- Names and Labels: Every interactive element (buttons, links, form fields, etc.) should have a clear, programmatically defined name so agents can understand its purpose.
- Tree Integrity: Roles, hierarchy, and parent-child relationships must be properly structured to ensure content is interpreted correctly.
- Visibility: Interactive content should be exposed within the accessibility tree and not hidden from assistive technologies or AI agents.”
A well-structured accessibility tree helps AI systems accurately understand page content, navigate interfaces, and complete tasks on behalf of users.
How to Access this audit in Chrome Dev Tools?
Enable the WebMCP Flag
- Open (or download) Chrome Canary. (Also make sure you have a recent version of the browser).
- Navigate to chrome://flags/#enable-webmcp-testing by pasting it into your URL bar.
- Change the dropdown menu setting to Enabled.
- Click Relaunch at the bottom right to restart the browser.
Run the Agentic Browsing Audit
Once the browser restarts, you can access the report directly via Chrome DevTools:
- Go to the website you want to test.
- Right-click anywhere on the page and select Inspect to open DevTools.
- Click the Lighthouse tab at the top of the DevTools panel. (If you do not see it, click the » arrows to view hidden tabs).
- In the Categories checklist, ensure the new Agentic Browsing checkbox is selected.
- Click Analyze page load to generate your report.
Our Take
Since this audit is coming from Google, we highly recommend that every website owner run this audit. If Google is prioritizing these three elements, they must be pretty critical to be discovered by the AI agents.
We’re still very early in understanding what works and doesn’t, pertaining to AI visibility, but as the AI tools are sharing more information, everything is starting to come together. For example, Google also shared their best practices on how to optimize a website for the Google AI experience.
For all of the marketers out there, it’s best to lean into this next frontier of organic visibility. Current referral traffic from AI sources might be small, but it’s growing month over month. Further, many of the recommendations for AI help with general website performance and traditional SEO.
