Observability and security platform company Datadog Inc. today unveiled more than 100 new capabilities at its annual DASH 2026 conference, headlined by a major expansion of its Bits AI agents that the company says can now run operations autonomously across the software development lifecycle.

The releases center on two pressures Datadog argues are reshaping enterprise technology: code being generated faster than humans can manage it and attackers using artificial intelligence to go after critical systems. The company is pitching deeper automation and broader visibility as the answer to both.

Bits AI, Datadog’s suite of agents for development, security and operational work, is the centerpiece. Previously focused on investigating the root cause of issues, it gains modules including Bits Detection, Agent Evals, Infrastructure, Code, Release, Data Analysis, Testing and Chat.

Together they let the agents detect, investigate and remediate problems by scanning infrastructure around the clock, recommending fixes and resolving them under predefined guardrails. The new Agent Eval feature also lets Bits AI debug and generate fixes for AI agents. Bits AI runs inside tools teams already use, including Slack and Anthropic PBC’s Claude.

On the security side, Datadog introduced AI Guard, which targets prompt injection and agent poisoning attacks. The product combines agent telemetry tracing with behavioral anomaly analysis to spot and block attacks that single prompt-and-response checks miss.

“With AI agents operating with elevated privileges, accessing sensitive data and communicating externally, a single malicious prompt hidden in an innocuous-appearing prompt can turn a well-intended agent into a malicious actor leaking sensitive information,” said Tim Knudsen, vice president of security products at Datadog.

The company also addressed the cost of soaring log volumes with Bring Your Own Cloud, which deploys its platform inside a customer’s own environment so data is processed and indexed in their cloud object storage rather than moved out. The approach is aimed at companies forced to choose between retaining data and absorbing the expense or deleting it and losing visibility.

Two further releases focus on agent oversight. Bits Agent Builder lets teams build custom AI agents inside Datadog to automate remediation, generate reports and enforce standards within customer-defined controls. Agent Console provides centralized monitoring for AI agents and agentic developer tools, including Claude Code, Cursor and GitHub Copilot.

The company framed the launches as a bet that operational control, not model quality alone, will separate winners in the AI era. “The companies that win on AI won’t just build better models, they’ll build operational control around them,” said co-founder and Chief Executive Olivier Pomel.

Datadog said it invests about 30% of revenue in research and development. The features were announced at DASH, where the company traditionally rolls out its largest product news of the year.

Photo: Datadog