Resolve AI’s SRE-1 model finds root causes of production alerts as AI agents go on call

SiliconANGLE
Read full story· 4 min
Resolve AI launched SRE-1, a 27B-parameter AI model trained to identify root causes of production alerts, improving investigation pass rates from 6.9% to 38.8%. Their new Primary On-call service uses AI agents to handle alerts autonomously, resolving up to 80% without human escalation.

What happened

Resolve AI introduced SRE-1, a 27-billion-parameter AI model designed to determine the root causes of production alerts. The company also launched Primary On-call, a service where AI agents autonomously handle production alerts before involving human engineers. SRE-1 was trained over 7.4 days using up to 512 GPUs and is only available through Resolve AI’s platform. Primary On-call is now generally available and has been tested by early customers including Salesforce.

Key details

  • SRE-1 has 27 billion parameters and was trained by 16 researchers.
  • Training used Snowflake’s Cortex Training service and reinforcement learning.
  • Investigation pass rates improved from 6.9% to 38.8% after training.
  • Primary On-call resolved up to 80% of paged alerts without human escalation.
  • Resolve AI has raised over $190 million and is valued at $1.5 billion.

Why it matters

Engineering teams at companies using Resolve AI’s platform can now automate the investigation and resolution of most production alerts, reducing the need for human intervention. This can lead to faster incident response and less disruption for on-call engineers, as reported by early users like Salesforce. Organizations adopting these tools may see lower operational costs and improved reliability in their production systems.


More in Business & Enterprise