Google Lighthouse now includes an Agentic Browsing category that audits really good AI agents are capable to navigate, understand, and enactment connected a site.
Google’s archiving notes that the Agentic Browsing class is experimental. Practically speaking, that intends you request to tally Lighthouse successful Chrome Canary, Google's experimental Chrome transmission that’s released daily.
The Agentic Browsing class runs 3 sets of checks:
- WebMCP: Validates your WebMCP integration — a projected web modular that lets a tract supply directions for AI agents to enactment upon
- Agent-centric accessibility: Reviews the portion of your accessibility character that matters astir for instrumentality interaction, including names and labels, character integrity, and visibility
- Stability and discoverability: Measures Cumulative Layout Shift (CLS) and checks for the beingness of an llms.txt file
Marie Haynes, Owner of Marie Haynes Consulting, shared an X post astir the announcement that pointed retired really absorbing it is that 1 of the checks centers connected an llms.txt file:
Per Google's documentation, the standards for the agentic web are still emerging. Instead of the acquainted 0–100 score, the Agentic Browsing class shows a ratio of really galore readiness checks your tract passes.
What the Agentic Browsing class signals for marketers
Agentic readiness has been easy to put disconnected until now, but the caller alteration to Lighthouse signals that isn’t feasible for overmuch longer.
Agentic protocols and devices are arriving fast. Just recently, Google has been building toward an agentic web with:
- AI capacity study successful Merchant Center: Tracks really AI-generated shopping experiences are performing
- AI Assistant transmission successful GA4: Attributes postulation from AI assistants arsenic its ain transmission successful Google Analytics
- Google-Agent personification agent: A dedicated crawler personality that signals erstwhile Google's AI agents are visiting and interacting pinch your site
- Universal Commerce Protocol: A modular that lets AI agents complete purchases crossed retailers without leaving Google's surfaces
- WebMCP: A projected web modular that, if adopted, would fto sites nonstop agents connected really to act
A Google-backed audit for agentic readiness provides hunt marketers pinch a reputable measurement to warrant their activity to stakeholders.
Google has besides said llms.txt isn't required for generative AI search. Its inclusion present suggests the record matters much erstwhile your tract has agent-specific elements, which is precisely what the Agentic Browsing study evaluates.
How to measurement visibility beyond Google's surfaces
For an supplier to enactment connected a website, it must first beryllium visible to AI platforms, including ChatGPT and Perplexity.
Semrush One combines some the AI Visibility Toolkit and the SEO Toolkit to fto you way your beingness crossed galore AI platforms and hunt engines successful 1 place.
Use the AI Visibility Toolkit to show citation wave and visibility trends crossed AI tools. And to spot which of your URLs get cited, the grounding queries that trigger those citations, and citation counts by platform.
Then corroborate you aren't accidentally blocking AI agents using Site Audit. Unintentionally blocking AI crawlers limits your AI visibility.
For larger and much analyzable organizations, Semrush Enterprise AIO reveals in-depth insights astir your stock of voice, visibility, and AI referral postulation crossed platforms.
