Grafana Cloud Adds AI Observability for Agent Conversations and Tool Calls

As companies deploy large language models in customer service, software development and automated workflows, AI agents’ decision paths and use of tools are emerging as new operational risks. Grafana Labs has therefore expanded AI Observability in Grafana Cloud to monitor conversations, model responses and calls to external tools, helping teams investigate performance, cost and security issues.

Grafana Labs’ latest features provide real-time visibility into AI agent conversation logs, tool calls and token usage. They can also evaluate output quality using an LLM-as-a-judge approach and support the OpenTelemetry standard. As of July 20, 2026, the available information did not disclose a formal launch date, pricing plans or investment amount, leaving the cost of enterprise adoption subject to further details from the company.