The pattern we kept ignoring
When you build software for a living, you start noticing the same problem wearing different costumes. The costume changes; the body underneath is always the same: valuable work gets created, and then quietly leaks away before anyone captures it.
We have now built two products around that one observation. Here is the story — and the lesson I wish we had internalized three years earlier.
Leak #1: billable time
Our first product started as an internal tool. We were a small consulting shop, and at the end of every month we reconstructed our hours from memory, calendar invites, and guilt.
The math was brutal. If each person under-reports just 20 minutes a day — a quick call here, a "real fast" review there — that is roughly 7 hours a month per person walking out the door unbilled. For a five-person team billing $100/hr, you are lighting about $3,500 on fire. Every month.
The fix was not a better spreadsheet. It was removing the memory step entirely: tracking time as it happens instead of reconstructing it later. That tool became FillTheTimesheet.
Leak #2: your best prompts
Fast-forward to the AI era. We watched our own team — marketing, sales, support, not engineers — get genuinely good at writing prompts for ChatGPT and Claude. Someone would craft a prompt that turned a 40-minute task into a 4-minute one...
...and then paste it into a Slack thread, where it died.
Two weeks later, three other people would reinvent a worse version of the same prompt. The institutional knowledge existed for exactly one session, then leaked away.
Same body, new costume. So we built PromptShip — a shared prompt library so a team's best prompts get captured once and reused by everyone, instead of getting lost in chat history.
The lesson
Here is the part I would tattoo on past-me: the most valuable problems are the ones so mundane that everyone has quietly accepted them.
Nobody files a feature request for "I keep forgetting my hours" or "our good prompts disappear." They absorb the loss as a cost of doing business. That acceptance is exactly where the opportunity hides.
If you are looking for something to build — or just something to fix in your own workflow — do not hunt for exotic problems. Look for the leaks everyone has stopped noticing: work that gets created but never captured, knowledge that lives in one person's head, value that is generated and then immediately discarded.
Plug one of those and you will rarely have to explain why it matters.
Key takeaways
- Valuable work leaks constantly: unbilled time, lost knowledge, discarded outputs.
- The best problems are mundane enough that people have stopped complaining about them.
- Capture-at-the-moment beats reconstruct-it-later, every time.
- You do not need a novel problem. You need an unaddressed one.