Groundcover AI Agent Now Acts Inside Slack, Linear, GitHub and Jira

Quick Facts

• Groundcover launched Remote MCP Connectors on June 22, letting its AI agent act inside Slack, Linear, Jira, and coding tools including Cursor, GitHub Copilot, and OpenAI Codex.
• Nearly 70% of groundcover customers are actively using MCP, and three in four are using at least one AI capability on live production data.
• The agent operates under each user’s own credentials and RBAC permissions, with every action logged in an audit trail and no code merged without human approval.

Groundcover has expanded its AI Agent Mode with Remote MCP Connectors, giving the agent the ability to reach outside the observability platform and act directly in the tools where engineering work happens. The connectors, which became fully available June 22, connect the agent to Slack, Linear, Jira, and third-party coding agents.

Before this release, the groundcover agent could investigate incidents, run queries, and surface root causes. But it could not act on findings outside its own platform. The new connectors close that gap.

The end-to-end workflow now looks like this: an alert fires, the agent reads the Slack thread where an engineer flagged a Postgres migration, pulls the linked Linear ticket, and traces the cause across logs, metrics, and Kubernetes events. When it lands on a likely fix, it hands off to a coding agent such as Cursor or OpenAI Codex, which opens a pull request. A human reviews and approves before anything merges.

The security model is built around the triggering user’s identity. The agent inherits the same role-based access controls and data scope as the person who launched it. If an engineer can only see one cluster, the agent working on that engineer’s behalf sees the same one cluster. Every API call lands in an audit trail.

Groundcover also exposes its own inbound MCP server, so external agents and tools a team has already built can query groundcover directly. The outbound connectors and inbound server are designed to work as two sides of the same integration layer.

The Slack integration goes further than routing alerts. According to groundcover, Slack is where offhand clues that explain a production problem tend to live, separate from any dashboard. The integration routes monitor notifications to the right channels and lets users call the agent into any existing thread.

Orr Benjamin, VP of Product at groundcover, said the company deliberately avoided treating the AI agent as a separate product. “You have some companies that are looking at their AI agent as a separate product entirely. That’s the polar opposite of what we want to do. We want to blend the experiences so that traditional observability and AI meet, and asking AI Mode feels like an extension of the same experience.”

The launch comes alongside an industry report groundcover released in May 2026. The report, based on a survey of 500 U.S. technology professionals conducted by Atomik Research, found that 49% of respondents attributed at least half of their observability costs to AI workloads. Only 22% said they were very confident their observability tools could detect real AI issues.

The same survey found that 87% of organizations have AI integrated into their observability workflows, but only 34% describe that integration as fully operational and trusted. More than a third cited data silos or fragmented tooling as a barrier to proactive decision-making.

Shahar Azulay, CEO and co-founder of groundcover, said the organizations that close the visibility gap now, especially around AI workloads, will move from reacting to incidents to shaping outcomes. Those that do not, he said, risk operating with an incomplete picture of their own systems.

Groundcover runs its Claude-backed Agent Mode through cloud-managed services the customer already uses, including AWS Bedrock, Google Vertex AI, and Azure Foundry. Customer data stays within their own cloud environment under the company’s bring-your-own-cloud model.

Read more: Groundcover lets AI agents work in Slack, Linear and GitHub with new connectors