Introducing Category Reviews that Update Themselves

Today we're releasing auto-load and auto-refresh in Sous: category reviews that update themselves. Point Sous at the place where your syndicated data exports already land. When a new period arrives, Sous loads the file, checks it, and rebuilds every report that depends on it. The next morning starts with one notification that your reports are current.

A category review that updates itself is a living report built on your syndicated data: when a new SPINS, Circana, or Nielsen file lands, the review's numbers, charts, and written narrative all refresh automatically, with no downloading, uploading, or rebuilding by anyone on the brand team.

Watch it happen:

What's new

Point Sous at your data once. A shared folder, a drive, or your warehouse: wherever your exports already land. That folder is the integration, so nothing about how your data arrives has to change, and there is no new upload step for anyone on the team to learn.

New periods load themselves. Sous sees the file the moment it lands, compares it against last period's, and loads it as a new version. History stays intact, so any prior period is still there exactly as it was.

If anything looks off, Sous stops and asks. A renamed column or a missing measure pauses the load before a single workbook is touched. You see what changed and decide, and nothing proceeds until you do.

Everything downstream rebuilds. Every workbook built on that data re-runs with new numbers, new charts, and a rewritten narrative, formatted in your brand's style.

You get one notification. When the refresh finishes, Sous tells you your reports are current.

Why this matters

The demo opens with the ritual: "Every four weeks, the same ritual at every CPG brand in America. The new data arrives. Someone downloads the file. Uploads it. Checks the columns. And then rebuilds every report sitting on top of it. A full day, gone. Every single period."

Depending on your feeds, that ritual runs every one to four weeks, and all of it is logistics. The analysis has not started yet, and a day is already gone. Auto-refresh returns that day to the decisions the data was supposed to inform.

Built to be trusted

The real risk with automation is silent failure: a refresh that quietly ships wrong numbers into a deck someone presents. Sous is built so that cannot happen quietly. The same provider rulebooks that check every Sous query against SPINS, Circana, and Nielsen conventions also define what a valid file looks like. A clean file refreshes on its own, and anything unexpected waits for a person. Our trust and accuracy chapter explains why we hold that line everywhere.

In practice

In the demo, a rep covering the Midwest has exports landing in one shared folder. Downstream sit her category review and a teammate's Kroger tracker. The file drops once and both rebuild: the full review, the analyses, the pitch, and the slides. Nobody on the team downloaded or uploaded anything.

This is a first concrete step toward the autonomous brand: routine data work runs itself, and people keep the judgment calls. For what that looks like inside a real brand, read the Fishwife story.

Getting started

Auto-load and auto-refresh are live now for every Sous workspace.

1. Sign in at platform.asksous.ai, open Data, and connect the drive or SharePoint folder where your exports land from the Connectors tab.
2. Open your dataset and turn on auto-load: pick the folder to watch, and Sous loads each new matching file as a new version of that dataset.
3. On the same dataset page, switch on auto-refresh for workbooks. When the next period drops, every review and tracker built on that data rebuilds, and the notification tells you when everything is current.

For the bigger picture on where this is going, start with the complete guide to agents in CPG.