Dynatrace wants its AI agents to fix problems instead of adding alerts

Dynatrace sells software that monitors companies’ apps and infrastructure and wants its AI agents to make engineers’ jobs easier. An agent that gets in the way is a bad agent and should be dropped, said Sean O’Dell, a principal product marketing manager at the company, when I talked with him at WeAreDevelopers on Sept. 26.

According to Sean, the company’s view is that AI apps should be monitored like any other app. They still run alongside older systems, including mainframes, and when something breaks, teams need to know what it cost and why it happened.

Dynatrace’s Bluebox agent, introduced in July, compares a team’s code with production data and proposes a fix as a pull request. Engineers can review it or let it run without them. The goal is fewer alerts, delivered in the tools developers already use, such as Claude and Codex, rather than in a separate dashboard.

Ops teams have rarely let automation change production on its own. In the ITIL era, an automated change came with layers of review. Dynatrace still tells customers to check a fix before approving it. “I don’t trust everything that comes out of my [chatbot] or my agent,” O’Dell said. “So why would you?”

The same rule applies to anything engineers make with AI’s help. If the work isn’t something you’d put your name on, he said, it’s slop, and it’s on the engineer to catch it before it ships.

He hopes teams will give agents more freedom over the next 12 to 18 months as they build confidence in them. That depends on the data behind the agents being accurate and available today. If it’s wrong now, it won’t be right in a year. If it’s right, the gains compound. Either way, he said, people should stay in control, and the work of building that trust has to start now.