Linear's Loops feature now reacts to project and cycle changes by editing docs and drafting Slack updates, taking over the coordination work a product team used to do by hand after every plan change.
A project's target date slips by a week, and the real work starts right after that decision gets made. Someone has to update the launch document so it matches reality, post a message explaining what changed and why, and make sure the two or three stakeholders who built their own plans around the old date actually see it. None of that is hard work. It is the kind of work that is easy to skip when the week is already full, and skipping it is usually how a team ends up finding out three weeks later that half the room was still planning around a date that no longer exists.
Linear's Loops feature update, shipped September 14, 2026, is built to do that follow-through automatically. Loops are recurring agent workflows inside Linear, and the September release lets them react to changes across a workspace: a project's status, ownership, target date, milestones, or labels; an initiative update or comment; a cycle starting or completing. When a change fires, a loop can edit the relevant Linear document to keep it aligned with the new plan, then draft and post a Slack update explaining what changed and who needs to act. Each run starts a conversation with Linear Agent that a person can review, add context to, or redirect afterward.
Loops themselves aren't new. Linear introduced them in July 2026 as a way to hand recurring team work to its agent: reviewing incoming issues and routing them to coding agents, turning meeting notes into follow-up work, keeping launch specs current on a schedule or in response to an event. What changed in September is the trigger surface and the output. Loops used to run on a schedule or off a narrower set of events. Now they respond to the kind of everyday project metadata changes that used to require a person to notice the change and decide who needed to hear about it, and they can write directly into the document and the Slack channel instead of just flagging that something happened.
The feature sits behind Linear's Business plan, which lists at $16 per user per month billed yearly. Loops isn't available on the Free plan or the $10 per user per month Basic plan, both of which cover core issue tracking without the agent workflow layer. On top of the seat price, Loops runs draw from Linear's AI credits, the same metered pool that powers coding sessions with Linear Agent. Linear covered early usage with $20 per seat in promotional credit when Loops launched in July, which expired August 20. The September post extends a new round of introductory Loops credit through the end of the year, which is worth noting for anyone budgeting past that date: the $16 seat price is not the full cost once the credits run out, and Linear hasn't published what usage looks like at scale without the subsidy.
For a team already paying for Linear's Business tier, this is a workflow it's already funding sitting a step closer to doing itself. For a team on Free or Basic, the calculation is different. Moving up to unlock Loops means paying for private teams, Triage Intelligence, Code Intelligence, and Linear Insights whether or not those matter, just to get the agent to write the update. That's a real jump for a small team where project dates rarely move and a Slack message takes two minutes to write by hand anyway.
The honest caveat is less about price and more about trust in the output. A September roundup of Linear's changes from Mean Tools makes a fair point about AI-drafted planning content generally: it can sound more settled and certain than the source material actually supports. A Slack update that reads as confident and final, drafted automatically the moment a target date changes, still needs a person to check that the underlying decision is actually decided before it goes out to a channel full of people who will act on it. Loops can be redirected after the fact, but that only helps if someone is watching the thread it opens.
What Linear is really testing here isn't whether an agent can write a passable status update. Plenty of tools can do that already. It's whether teams will trust an agent to decide, on its own, when a status update needs writing at all, which project changes are the kind that deserve a message and which aren't. That's a bigger ask than automating the typing, and it's the part worth watching as more of these updates start showing up in Slack channels without a person's name attached to the sending.