Free Idea: The Drive That Cleans Itself
Free idea. I'm not building it, but I wish it existed every day. Someone please take it.
Nobody has a clean shared drive. Not the Fortune 100s I've worked with, not the startups, not myself on my own desktop. I don't think I've seen one. Ever.
And you know exactly how it happens, because you probably did it this morning. You're slammed. You skip the naming convention. You drop the file in the first folder that sort of fits and you move on with your life. Multiply that by a few dozen people over a few months, and eventually someone sends a desperate message: "can we PLEASE clean up the drive." A cleanup happens. Holds for about nine days. Rinse, repeat.
Picture this instead
You hit save. Something quietly reads the file, works out what it is, names it properly, tags it, and puts it where it actually belongs. If it's not sure, you get one popup with one question. "Looks like this could live in Strategy or Analytics. Pick one." You click. It disappears. Two seconds and you never think about that file again.
That's the whole pitch. Plugs into Drive, SharePoint, Box, whatever you're already on.
The pieces that already exist, and why nobody uses them
One caveat, because I actually went and looked. Google shipped "Organize My Files" in Drive a couple of months ago. Microsoft's SharePoint Premium has been auto-classifying documents for years. So pieces of this already exist.
What surprised me more is that I'd never heard of either one until this week, and I can't name a single person who uses them. That tells you something on its own. Google's only fires when you go hunting for a cleanup, and it doesn't rename anything. Microsoft's needs you to train it per document type. Neither one knows who you are or what you're working on.
They also only work on their own turf. Google's won't touch SharePoint, Microsoft's won't touch Drive, and neither one is ever going to look at your desktop, which in my case badly needs it. Most of us are spread across all three.
What the real version has to do
The version I want lives on your machine and reads whatever you've got, which neither of them can do. Three things make it work.
It acts at save time, not at cleanup time. The mess happens one file at a time. The fix has to happen one file at a time too, in the two seconds you're already there.
It asks at most one question. A tool that asks three questions per file is a tool people turn off by Wednesday.
It knows your work. Your projects, your clients, what you were in last week. That context is what turns a guess into a decision, and it's the part the cloud tools structurally can't have.
Somebody go build it. I'll be your first user. And I'll bring you the messiest drive you've ever seen.