AI agents and your website

What AI agents are

“AI agents are autonomous systems that can perform tasks on behalf of people, such as booking a reservation or comparing product specifications.” Browser agents may read a website by analysing screenshots, inspecting the DOM and interpreting the accessibility tree.

Officially documented · Google Search Central

Google presents this as optional: worth reviewing “if this is something that’s relevant to your business and you have extra time”.

How agents see a page

“Agents can view your website in 3 primary ways: screenshots, raw HTML, and the accessibility tree.” Modern agents combine them. The DOM and accessibility tree give a structured list of interactive elements. Agents cross-reference that list with a visual rendering for layout and grouping.

Officially documented · web.dev (Google)

Practices

The web.dev guide recommends:

• Avoid ghost elements: transparent overlays that hide interactive elements.
• Use semantic HTML: <button> and <a> over modified <div> and <span>. Otherwise add the right role and tabindex, and link <label> to inputs with for.
• Make interactive elements visible and “larger than 8 square pixels”.
• Keep the layout stable, so actions stay in predictable places.

Officially documented · web.dev (Google)

“Everything we suggest to make a site ‘agent-ready’ also makes sites better for humans.”

Officially documented · web.dev (Google)

Not established

• Whether agent-friendliness affects how often AI search chooses or cites a site. These practices are about agents completing tasks on a site.
• Emerging protocols such as WebMCP and the Universal Commerce Protocol are proposals or early-stage; behaviour may change.

How this wiki does it

Pages are server-rendered semantic HTML with real links and buttons. They pass an automated WCAG 2.1 AA audit, which checks the same accessibility tree an agent reads.