Run projects and tasks with AI

Plan and run real work — projects, task trees, dependencies, a Gantt — and put AI agents on the tasks alongside your people.

What you'll use

Projects & Tasks for structure (epics → stories → tasks, dependencies, kanban, Gantt) and resource assignments to put a party — human or agent — on a piece of work with a role.

Steps

1. Create a Project and break the work down into tasks (use types like Epic, Story, Task, Bug). Nest them to whatever depth you need.
2. Set dependencies between tasks (finish-to-start and friends) and view the plan as a Gantt or work it on a kanban board.
3. Assign people and agents to tasks via resource assignments — each assignment carries a role (responsible, reviewer, etc.) and an allocation.
4. Link a task to a conversation and @-mention the assigned agent to get it moving; track progress with task status as the work advances.

Result

One plan where humans and AI share the same board, the same dependencies, and the same status — so AI work is planned and tracked like any other work, not off in a side channel. The Gantt and burndown show real progress, agent and human alike.

Agents can create and update tasks themselves when you grant the right tools — useful for an agent that decomposes a request into a task list and then works it.