Shared AI Workspace Puts Incident Investigations in One Place
Agent Investigations lets outside agents such as Claude Code and Codex join production debugging, a shipped feature for customers with AI enabled.
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 work production incidents together. The feature is available to groundcover customers whose backends have AI enabled, so it is shipped rather than announced.
Why it matters
Several engineers often point their own agents at the same production issue, each working in isolation, so conclusions end up in private chats, tickets, or Slack messages out of view of the team. Agent Investigations moves that work into one shared, auditable space, reflecting a broader shift toward multi-agent debugging of live systems.
The figures reported are groundcover's total funding of $160 million and its $100 million Series C round, which are financing metrics rather than technical performance measures. No parameter counts, latency numbers, or accuracy benchmarks for the agents were disclosed.
The players
groundcover
An observability company whose platform collects logs, metrics, traces, and Kubernetes data, and which now offers AI agents for production incident investigation.
One Peak
An investment firm that led groundcover's $100 million Series C round in July 2026.
Amazon Web Services Inc.
A cloud computing company whose re:Invent conference in Las Vegas is where groundcover plans to demonstrate Agent Investigations.
The details
Investigations start automatically when groundcover opens a monitor issue, or engineers can launch one manually from the interface or a Slack button. The Investigator agent builds a graph of questions and hypotheses, then tests each against logs, metrics, traces, and Kubernetes data. A separate Moderator agent is the only agent permitted to close a hypothesis and sends back any claim that lacks supporting findings, with status posted to the Slack thread tied to the original alert.
Timeline
One Peak led groundcover's $100 million Series C round in July 2026.
groundcover launched agent-to-agent observability tools in October 2026.
Agent Investigations launched in November 2026, and groundcover plans to demonstrate it at re:Invent from Nov. 30 to Dec. 4, 2026.
The Tech Race
Agent Investigations positions groundcover among observability vendors building agent-native incident response rather than dashboards alone. Its Model Context Protocol server lets rival agents such as Claude Code and Codex plug into the same production data, placing it in a contest over which platform becomes the default context layer for AI-driven operations.
Engineers at groundcover customers whose backends have AI enabled can use Agent Investigations now, with investigations opening automatically from monitor issues. Teams that already run their own coding agents can connect them through the Model Context Protocol server, while the source gives no pricing details.
The takeaway
The trajectory is toward shared, auditable AI investigation of live systems rather than isolated per-engineer agents. Watch the live demonstration at re:Invent from Nov. 30 to Dec. 4, 2026, for evidence of how the Moderator agent performs under real production incidents.
What happens next
groundcover is scheduled 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 these tools are changing engineering work, explore our Artificial Intelligence coverage.

