AI agents and automated code drive demand for new observability tools

The rapid integration of AI agents and AI-generated code into software development is driving an urgent demand for new observability and telemetry capabilities. As AI-generated code is projected to assist with 65% of shipped code by 2027, companies are seeking ways to monitor autonomous systems that operate too quickly for human oversight.

To address this, Hud has integrated with ClickHouse’s ClickStack to create a runtime feedback loop. This integration allows engineering teams to assess AI-generated changes before deployment and automatically investigate or roll back code if production behavior deviates. Monday.com is currently an early adopter of this stack.

However, the rise of AI agents introduces significant strategic and security risks. Experts warn that the telemetry data generated by AI—such as prompts and tool calls—could be used by model providers to capture institutional knowledge and gain a competitive advantage over their enterprise customers. Additionally, security teams face new challenges, including the need for millisecond-level automated responses to combat AI-driven attacks and the risk of hallucinating security agents causing network damage.