What changed
Resolve AI Inc. announced the launch of SRE-1, a new model trained to identify the root causes of production alerts, and Primary On-call, a service that deploys its AI agents to handle incidents before human engineers are involved. The model was developed using open weights and trained by a team of 16 AI researchers over 7.4 days on up to 512 graphics processing units.
What this means in practice
SRE-1 is designed to improve incident resolution by analyzing production alerts and identifying their root causes. According to Resolve AI, the model increased the pass rate for a single investigation from 6.9% to 38.8% when tested on incidents before and after training. The company also reported that its complete system solved about twice as many incidents as general-purpose coding agents when benchmarked on 178 production alert investigations.
Primary On-call puts this model into action by investigating alerts across telemetry, infrastructure, code, and operational knowledge. The AI agent can close alerts, mitigate issues where permitted, or escalate to a human.
It also links related alerts into a shared view of each underlying problem. For teams hesitant to fully automate, Resolve AI offers a shadow mode that records the agent’s actions without executing them.
Limitations and availability
Access to SRE-1 is limited to Resolve AI’s own platform. Primary On-call is now generally available. The company claims up to 80% of paged alerts in production are handled by Primary On-call without escalation, with customers reporting a tenfold improvement in mean time to resolution and root-cause accuracy of up to 95%. Salesforce Inc. is among the early users, with reports that the agent closed over 70% of the CRM Analytics team’s paging alerts during testing.
