Google draws agentic line
Google clarified the technical boundary between user‑triggered AI agents and search crawling systems like Googlebot, a move that signals growing engineering and compliance specialization around agentic systems (marktechpost.com).
Google added a new user agent string, "Google‑Agent", to its official list of user‑triggered fetchers on March 20, 2026. (developers.google.com) Google describes the Google‑Agent user agent as the identifier used by agents hosted on Google infrastructure to navigate websites and perform actions on behalf of users. (developers.google.com) Google published the associated IP ranges and verification guidance in its crawling infrastructure files (including user‑triggered‑agents.json) and recommends verifying requests by matching accessing IPs or reverse DNS patterns. (developers.google.com) The company’s documentation notes that user‑triggered fetchers generally ignore robots.txt rules, meaning Google‑Agent requests may bypass robots.txt constraints that govern traditional crawlers. (developers.google.com) Google said the Google‑Agent user agent is in a gradual rollout and will be deployed “over the next few weeks,” indicating phased activation rather than an instantaneous switch. (seroundtable.com) Google’s Project Mariner — the DeepMind‑led web‑browsing agent that automates multi‑step tasks — is explicitly cited by multiple outlets as an example of an agentic system that will surface as Google‑Agent traffic. (searchengineland.com)
