Linear launches Agent: AI built natively into project management

Linear released Linear Agent on March 24, 2026 — an AI assistant built directly into the Linear task management platform. Unlike bolt-on AI features that operate on isolated inputs, Linear Agent has access to the full workspace context: roadmap, backlog, issues, threads, customer feedback, and linked code repositories.

The interface works through a chat panel (keyboard shortcut Cmd/Ctrl+J in the desktop app, or via Slack and Microsoft Teams). A PM can write a natural-language prompt — “create a sprint from these five customer requests” or “write a spec for this feature based on the feedback thread” — and the agent creates issues, projects, and documents accordingly.

Two additional capabilities are in public beta: Skills (reusable saved workflows triggered on demand) and Automations (workflows that fire automatically when issues enter triage). A third capability, Code Intelligence, is coming later and will allow the agent to answer questions about which systems own which functionality and to diagnose bugs by reading the codebase alongside the issue.

Linear notes that 25% of new issues created in the platform are now agent-created, a figure that was 5x lower three months ago. As issue creation shifts toward automated agents, the PM role in this workflow shifts from writing and triaging tickets to reviewing and exercising judgment on agent-produced output. Linear Agent is explicitly positioned as the aggregation layer that converts scattered context — Slack threads, feedback sessions, backlogs — into structured work items.

The chat interface is free during beta on all plans. Automations and Code Intelligence will require Business or Enterprise plans at general availability. High-volume compute usage is expected to shift to usage-based pricing.

For PMs managing complex roadmaps across multiple stakeholders, the value proposition is reducing the coordination tax: less time writing tickets, more time deciding which work matters.