Datadog Inc. (NASDAQ: DDOG) announced the general availability of Bits Agent Builder, a tool that enables users to create custom AI agents for automated workflow management and alert response.

The new feature allows teams to build AI agents that analyze observability data and execute responses without requiring hardcoded logic for every scenario. These agents can interpret data from Datadog's monitoring platform and third-party sources to automatically investigate issues and trigger remediation actions.

Bits Agent Builder integrates with Datadog's Action Catalog, which contains over 2,000 prebuilt actions connecting to external services including Kubernetes, cloud platforms, CI/CD tools, and security systems. Users can define agent goals using natural language and specify which data sources and tools each agent can access.

The platform supports workflow automation across several use cases. For application error investigations, agents can automatically search logs and real user monitoring data when triggered by support requests, then compile structured reports with evidence links. For security vulnerabilities, agents can create incident cases, quarantine affected systems, and notify engineering teams when critical issues are detected.

The tool also works with Datadog's Internal Developer Portal Scorecards to automate service documentation. When scorecards identify missing documentation, agents can generate initial README files using service metadata and telemetry data, then create pull requests for team review.

Multiple agents can be chained together in workflows, with conditional triggers based on specific situations. An investigation agent might analyze system data to identify root causes, then activate a remediation agent if the suggested fix matches known patterns.

Bits Agent Builder is accessible through the Actions section of Datadog accounts. The company offers a 14-day free trial for new users.