groundcover Launches Shared AI Agent Investigations Workspace
The new tool lets AI agents and engineers work production incidents together in one shared workspace.
Updated on Nov. 19, 2026 in Artificial Intelligence

groundcover has launched Agent Investigations, a shared workspace where its AI agents, team coding agents, and on-call engineers can work production incidents together. Outside agents, including Claude Code and Codex, can join through groundcover's Model Context Protocol server.
Why it matters
Multiple engineers often point their own agents at the same production issue in isolation, and their conclusions can end up in private chats, tickets, or Slack messages that the rest of the team never sees. A shared workspace aims to keep those findings in one place.
Investigations start automatically once a monitor issue opens, and the Investigator agent tests each hypothesis against groundcover's logs, metrics, traces, and Kubernetes data. The Moderator agent is the only agent permitted to close a hypothesis, and the source does not describe performance benchmarks for the system.
The players
groundcover
groundcover is a company that builds observability tools for cloud and Kubernetes environments and offers AI-enabled monitoring for its customers' backends.
Amazon Web Services Inc.
Amazon Web Services Inc. is the cloud computing arm of Amazon, and its re:Invent conference is one of the largest annual cloud industry gatherings.
The details
The Investigator agent builds a graph of questions and hypotheses before testing each against groundcover's data, while the Moderator agent sends back any claim lacking supporting findings. Hypothesis status and assessments post to the Slack thread tied to the original alert, and engineers can also launch an investigation manually from the groundcover interface or a Slack button.
Timeline
July 2026: One Peak led groundcover's $100 million Series C round.
October 2026: groundcover launched agent-to-agent observability tools.
Nov. 30 to Dec. 4, 2026: groundcover plans to demonstrate Agent Investigations at Amazon Web Services' re:Invent conference in Las Vegas.
The Tech Race
Agent Investigations builds on the Model Context Protocol, an open standard for connecting AI agents to outside tools, to let third-party agents such as Claude Code and Codex work inside groundcover's production data. The launch places groundcover among observability vendors competing to become the shared workspace where AI agents handle live incidents.
Engineers responding to production incidents can have their own agents and teammates' agents work from one shared record instead of scattered private chats and tickets. Access is limited to groundcover customers whose backends have AI enabled.
The takeaway
Keeping AI agent findings in a shared, visible record may matter as much as the agents' own accuracy when teams respond to live incidents. Engineers evaluating such tools should check whether their monitoring backend supports AI features first.
What happens next
groundcover plans to demonstrate Agent Investigations at Amazon Web Services' re:Invent conference in Las Vegas from Nov. 30 to Dec. 4, 2026.
Further reading
For broader context on how autonomous tools are reshaping engineering operations, see Artificial Intelligence coverage.



