Nobody wants to upload a file to a product they don't trust yet

For a long time our empty state was a box that said: drop a document here.

Reasonable. It's a tool for reading long documents, so of course it needs a document. But watch what that actually asks of a first-time visitor. Go and find a file. Make sure it's the right one. Make sure it's long enough to be worth the trouble. Upload it. Wait. Only then do you find out whether the thing is any good.

Every step there is a place to leave, and the person hasn't seen a single reason to stay yet. We were asking for effort before we'd earned any.

So we built the opposite: a shelf of documents already loaded and ready to use. Court rulings, SEC filings, university course notes, startup legal templates. You pick one and start asking questions immediately. No file, no upload, no account required to look.

The bet is that the first thirty seconds should demonstrate value rather than request it, and that people will happily bring their own material afterwards. Too early for me to claim it worked. Follows are up, and whether that turns into people uploading their own documents is the number I'm actually watching.

What I didn't expect is how much of the work turned out to be curation rather than engineering. Which documents you put on that shelf is a positioning decision. Put finance filings there and you tell every visitor who this is for. Put in a textbook and you say something different. We're still arguing about the mix internally.

So, for anyone whose product needs an input before it does anything, a spreadsheet, a repo, a calendar, a dataset: did you solve the cold start with sample content, and did people ever graduate from the samples to their own stuff? I've heard the failure mode described as demo mode, where users play with the examples, never convert, and your activation metric quietly means nothing. I'd like to know if that's real before we lean harder on it.