New York based Cloud monitoring as a Service firm Datadog has unveiled a slew of new AI-powered observability and security technologies at the company’s annual DASH event.
“AI has created new operational challenges where code development has outpaced human-scale management and malicious actors now use AI to attack critical systems. But AI didn’t create this complexity — it accelerated what was already there. The companies that win on AI won’t just build better models, they’ll build operational control around them,” said Olivier Pomel, co-founder and CEO at Datadog.
“We consistently invest about 30 per cent of revenue into R&D, which is why we are able to deeply understand and solve the problems our customers face every day in managing operational complexity,” said Alexis Lê-Quôc, co-founder and CTO at Datadog. “At DASH, we launched 100+ capabilities unified around one goal: giving customers the visibility they need to find and fix the issues that matter most, the moment they matter.”
Key new products include:
Bits AI Becomes Truly Autonomous for Both Incident and Development Actions
Bits AI is Datadog’s suite of agents built to automate development, security and operational workflows.
“To date, Datadog’s Bits AI has focused on investigating the root cause of issues. Now — with Bits Detection, Agent Evals, Infrastructure, Code, Release, Data Analysis, Testing and Chat — Bits AI is capable of truly autonomous operations, becoming a reliable teammate that operates across every stage of the production lifecycle and development loop,” said Lê-Quôc.
Following the updates, Bits AI automatically detects, investigates and remediates issues by scanning infrastructure around the clock to uncover issues, recommend fixes and resolve them, using robust, pre-defined guardrails. In addition, with its new Agent Eval capabilities, Bits AI can debug and generate fixes for AI agents.
Bits AI also operates across the development loop by following each release and pull request from code change, staging and rollout through to production.
Protect AI Agents With AI Guard
“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 — costing millions in reputational damage and data loss. But attackers have learned to hide agent poisoning using subtle instructions only detectable with a deep understanding across multiple steps of the agent’s behaviour. Datadog’s new AI Guard uses a unique combination of deep agent telemetry tracing and AI-native stateful behavioural anomaly analysis to detect and block AI agent attacks otherwise missed by stateless prompt-and-response evaluation,” said Tim Knudsen, Vice President of Security Products at Datadog.
Bring Your Own Cloud
Datadog said that as AI volumes increased, the number of logs generated was growing exponentially. “This often forces companies to make an uncomfortable decision: keep the data and absorb the cost, or delete it and risk losing visibility,” it said. “Datadog is addressing this concern directly with Bring Your Own Cloud, which deploys the Datadog platform into a customer’s own environment so that data is processed and indexed in their cloud object storage.”
Create and Monitor Agents Within Datadog
Using the firm’s Bits Agent Builder, teams can create custom AI agents inside Datadog that automate remediation and operational workflows — resolving incidents, generating tailored reports and enforcing standards across environments. To help teams understand the business value of their agents, Datadog also announced Agent Console. This product provides centralised monitoring for AI agents and agentic developer tools like Claude Code, Cursor and GitHub Copilot.
“Agent Console provides the needed visibility to answer the key questions for users about the heaviest adopters of agents, the tasks that agents perform best and where they struggle, and how the work produced by agents correlates with spend,” said Lê-Quôc.
