Automation

Resolve AI launches SRE-1 model to automate incident response

The 27-billion-parameter model handles up to 80% of production alerts without human escalation, marking a shift toward AI-first on-call operations.

Omega Editorial · November 19, 2026 · 3 min read

Key takeaways

  • Resolve AI's SRE-1 is a 27-billion-parameter model trained specifically for production incident diagnosis, increasing single-investigation pass rates from 6.9% to 38.8%.
  • The Primary On-call service handles up to 80% of production alerts without human escalation, with customers reporting tenfold improvements in mean time to resolution.
  • Salesforce closed more than 70% of CRM Analytics paging alerts during initial testing without waking engineers.
  • The company has raised over $190 million and reached a $1.5 billion valuation, with customers including Coinbase, DoorDash, Nasdaq and SAP.
  • SRE-1 was built using Snowflake's Cortex Training service and trained over 7.4 days on up to 512 GPUs.

Resolve AI has released SRE-1, a specialized model designed to diagnose production incidents, alongside Primary On-call, a service that positions AI agents as the first responder to system alerts before escalating to human engineers.

The company built SRE-1 as a 27-billion-parameter model trained specifically for site reliability engineering tasks. According to details first reported by SiliconANGLE, a team of 16 researchers completed the final training run over 7.4 days using up to 512 GPUs. The model relies on reinforcement learning and was developed using Snowflake's Cortex Training service.

Performance benchmarks

Resolve AI measured SRE-1's effectiveness on the same set of incidents before and after training. The pass rate for a single investigation increased from 6.9% to 38.8%. In testing across 49 difficult production cases, each run five times, an ensemble of five SRE-1 investigations matched the performance of leading frontier models at lower cost.

The company also benchmarked its complete system on 178 production alert investigations, where it solved approximately twice as many incidents as general-purpose coding agents.

Primary On-call in production

Primary On-call applies the SRE-1 model to live incident response. When an alert triggers, the agent investigates across telemetry data, infrastructure, code and operational knowledge. Based on its findings, it can close the alert, implement mitigation where team policies allow, or escalate to a human engineer. After taking action, the agent monitors the fix against production baselines until recovery stabilizes.

The service also correlates related alerts and investigations into a unified view of each underlying problem. Teams concerned about immediate automation can use a shadow mode that records what action the agent would have taken without executing it.

Resolve AI reports that Primary On-call handles up to 80% of paged alerts without escalation. Customers have seen mean time to resolution improve by a factor of ten, with root-cause accuracy reaching 95%.

Early adoption

Salesforce is among the first organizations using the system. Bhupesh Patel, a software engineering principal architect at Salesforce, said that during initial dark launch testing, the agent closed more than 70% of the CRM Analytics team's paging alerts without waking an engineer.

"No engineering team can keep up with how fast code and patches land in production," said Resolve AI Chief Executive Spiros Xanthos. "The answer is agents that are accurate and fast enough to sit on the critical path, and that takes a model built for production."

Why it matters

Production incident response has remained largely manual even as AI transforms software development, creating an operational bottleneck for engineering teams managing increasingly complex systems. By automating root cause analysis and initial remediation, specialized models like SRE-1 could reduce on-call burden and accelerate incident resolution at a time when deployment velocity continues to increase. The shift from shadow mode to primary responder represents a significant expansion of AI's role in critical operational workflows.

Availability and backing

Primary On-call is now generally available. Access to the SRE-1 model remains limited to Resolve AI's platform. The company has raised more than $190 million in venture funding, including a $40 million Series A extension in April led by DST Global and Salesforce Ventures that valued the company at $1.5 billion. Customers include Coinbase, DoorDash, Figma, MSCI, Nasdaq, SAP and Zscaler.

These details were first reported by SiliconANGLE.

  • #site reliability engineering
  • #incident response
  • #ai agents
  • #devops automation
  • #resolve ai
  • #production monitoring

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Automation