Always-on agents run production without the on-call tax — Justin Smith, Resolve AI
Justin Smith from Resolve AI explains how agents—AI programs that run automatically—can monitor production environments without requiring developers to be on call around the clock.
When a new code version is released, the agent can read what changed, determine which monitoring metrics are relevant to that specific change, and create a customized control plan. The agent can also decide when to check again—perhaps in an hour if it's a type of error that appears intermittently, or in three days to verify everything remains stable. Smith emphasizes that approximately 70 percent of a developer's time goes to operating code rather than writing it, and that new AI tools have actually made this worse by increasing the volume of changes that must be monitored. Resolve agents work by answering three questions: when they should run, how they run (isolated in the cloud so it doesn't matter if you close your laptop), and how they know what to do.
Vibekollen prepared this summary with AI from the original publication. The content belongs to AI Engineer.
