How Magnolia keeps its research and product teams in sync with Slate

The challenge

Magnolia is a company full of people who build intelligent tools for a living. The irony was that their own work was anything but organized.

Research moved fast and loose, the way research does. Product needed structure and predictability. The two halves of the company were pulling in different directions, and the work connecting them lived in scattered documents that went stale almost as soon as they were written.

A model breakthrough in research would take days to reach the product team. A shift in product priorities would reach research even slower. By the time everyone was aligned, the ground had often moved again.

We build AI that keeps companies organized, and yet our own research and product teams could barely stay on the same page.

Mei Lin Chen - VP of Research, Magnolia

For a company moving as fast as Magnolia, that lag between knowing something and acting on it was the most expensive problem they had.

Why Slate

Magnolia moved to Slate to put research and product in one workspace, and they were drawn in by something most tools didn't have: an intelligence layer of its own.

As a team that builds AI, we are hard to impress. Atlas was the first time a tool's intelligence actually felt useful rather than bolted on.

Coming from a company of AI researchers, that was real praise. Atlas didn't just store the company's work, it understood it. It connected research notes to the product plans that depended on them, surfaced what had changed, and flagged where the two teams were drifting apart before the gap became a problem.

For a team that spent its days judging AI, finding one they actually wanted to use said something.

How they use it

Today, research and product at Magnolia share a single workspace, with Atlas working quietly in the middle of it.

Research documents its work in Slate as it goes, and those notes link directly to the product initiatives they affect. When a finding lands, the people who need it see it in context rather than hearing about it a week later.

Atlas carries much of the load of keeping two fast teams aligned. It surfaces what moved across the company overnight, connects related work that lives in different corners, and points out where decisions are waiting.

Atlas connects a research finding to the product work it affects, so nothing important sits unnoticed for days anymore.

Product uses Slate to tie its roadmap to the research underneath it, so plans rest on what the team actually knows rather than what it knew last month. When the research shifts, the plans visibly shift with it.

What changed

The clearest change was how fast knowledge traveled.

What used to take days to cross the company now moves the moment it happens.

The lag between a discovery and the action it should trigger mostly closed. Research reached product while it was still fresh, and product steered research while there was still time to steer it.

The two teams stopped drifting, because Atlas made the drift visible the moment it started. And the scattered, stale documents that used to hold the company's shared knowledge gave way to one workspace that stayed current on its own.

The result

Magnolia now runs research and product on Slate, with Atlas holding the two together.

The company moves faster because what one team learns, the other knows almost immediately. Decisions rest on current information instead of last month's. And the people building Magnolia's models spend their time on the work itself, rather than on the endless effort of keeping everyone in the loop.

We build AI to keep teams in sync, and Slate is the first tool we have used internally that genuinely does it for us.

For a company that teaches machines to make sense of scattered information, running on a tool that does the same for their own work turned out to be a natural fit.