Predicting Your Cash Crunch Was Never the Hard Part

A founder I work with, $9M in revenue, bought a working capital tool this spring. It promised to predict her cash position 90 days out, and it delivered. In April the dashboard told her late June would be tight. She showed it to me, pleased with the precision. I asked what she planned to do about it. She didn't have an answer.

She already knew June would be tight. Her controller knew. The two salespeople with big invoices outstanding knew. The software drew a sharper line under a problem nobody was acting on.

The forecast was never the bottleneck

Working capital optimization is the single most popular place finance leaders want to point AI right now. In one 2026 survey, 46% named it the top area where they expect to apply advanced capabilities. The pitch writes itself: an agent that predicts working capital needs, flags the shortfall, sees around the corner.

Here's what that pitch quietly assumes. It assumes you don't already know.

Most founders running a growing company can feel a crunch coming weeks out. They know the big renewal slipped. They know two customers stretched from net 30 to net 55. Better prediction sharpens a picture they've mostly already drawn in their head. It's a nicer demo than it is a fix.

Think of working capital work in three steps. See the problem. Decide what to do. Act. The tools sell "see." Founders think they're buying "act."

What actually moves the number

The cash conversion cycle isn't computed. It's negotiated and executed. Your DSO is the sum of when invoices go out, when dunning emails fire, and which customers you're willing to push. None of those are forecasting problems.

Take a $6M services firm carrying $1M in receivables. Cutting DSO from 55 days to 42 frees roughly $215K in cash. The agent can tell you that gap exists. It can even draft the reminder. What it can't do is decide whether you're willing to send a firm note to the client who's also your biggest reference, or whether to delay a vendor PO another week, or which of three late accounts gets the call today.

Prediction is the demo. The value is in the boring execution, and almost nobody sells that, because execution means touching your customer relationships.

The invoice that goes out the day a milestone clears instead of three weeks later. The dunning sequence that starts at day 31, not day 60. The deposit you ask for on a new logo before kickoff. These move working capital more than any 90-day forecast, and they're decisions, not outputs.

Buy the part you're actually missing

Before you sign for another tool, ask which of the three steps you're weak at. Most companies aren't weak at seeing. They're weak at the follow-through after the dashboard turns red, because no one owns it and every action requires a slightly awkward conversation.

If that's you, a sharper forecast won't help. A person whose job is to chase the invoice on day 31 will. Sometimes that person can be an agent, but only once you've decided the rules it's allowed to act on. Decide those first.

A model that predicts the crunch and a team that prevents it are not the same purchase. Know which one you're writing the check for.