Sentinel AI Analyst now closes 96% of tier-1 alerts autonomously across the Guardian fleet
Twelve months after general availability, Guardian's reasoning engine is resolving the overwhelming majority of first-line security alerts without human involvement — while median containment time across all customers fell to 8.4 seconds.
The figures come from anonymised production telemetry covering 6,800 organisations in 142 countries. Guardian is publishing the underlying methodology alongside the announcement, including how autonomous closures are audited, how false-positive rates are calculated, and the conditions under which the platform escalates to a human rather than acting.