When the procurement mess shows up in the cash account
A treasurer carrying two point one million in unreconciled cash every Monday morning is not unusual at mid-market scale. The number rarely comes from treasury. It comes from procurement. A note on why we built Calabash buyer-side first.
A treasury head I spoke with last year had a number stuck in her head. Two point one million. That was the amount of unreconciled cash she was carrying every Monday morning, sitting in clearing accounts she could not deploy because she did not know which supplier the funds were tagged to or whether the invoice they were paying had actually been received.
That number did not come from treasury. It came from procurement.
The team had been operating on emails and a shared folder for two years. POs got raised, GRNs got recorded when someone remembered, invoices arrived in a generic inbox, payments went out on Friday because that is when the bank window suited operations. Somewhere in that chain the matching broke, the supplier files did not line up with the bank files, and the treasurer was left holding the gap.
She is not the only one. The pattern shows up in almost every mid-market finance team we talk to. Procurement looks like its own function, with its own software, its own people. Then a misrouted PO or a missing GRN becomes a payment that cannot be cleared, an FX position that cannot be closed, a cash forecast that wobbles by a six figure number every week.
The handoff is the failure point
When we started Calabash we spent a lot of time arguing about where to begin. Buyer side or supplier side. Treasury or AP. Approvals or payments. We picked buyer-side AP-and-treasury together for a reason, and that reason is the handoff.
The handoff between procurement and finance is where most of the pain in a mid-market business hides. Two systems, two teams, two sets of master data, and the supplier sits between them wondering why their payment is late. Every other treasury problem you can name (cash visibility, supplier risk, payment fraud, FX timing) is downstream of that handoff being broken.
Fixing the handoff is not glamorous work. There is no AI feature you can slap on it. What there is, is a model. Invoices belong to a supplier business partner. The supplier business partner has bank accounts that are vetted by treasury, separately from the bank accounts the supplier themselves submitted during onboarding. Workflow steps live as rows that anyone can audit. Approvals carry forward. Reconciliation is the byproduct of getting those primitives right, not a separate ritual that happens at month end.
What the treasurer actually needs
If you ask a treasurer what they want from a procurement-adjacent system, the answer is rarely about features. They want to know two things on any given Monday morning. What did we owe last Friday and did it go out. What are we going to owe this Friday and is the cash there.
Everything else (approval flows, supplier portals, three way match) earns its keep by making those two answers reliable.
That framing is the reason buyer-side Calabash starts where it starts. We are not trying to replace procurement software, and we are not trying to be a payments platform. We are trying to make sure the data that lands on the treasurer's desk on Monday is the data she can trust on Monday, not the data she has to rebuild on Tuesday.
The next two weeks
The next post in this series is about why contract compliance, which sounds like a procurement problem, is actually a treasury time sink. After that, hidden costs in AP workflows, the manual processes we are trying to eliminate, and the specific architectural decisions that came out of a year of conversations like the one above.
If you are a treasurer or an AP lead at a mid-market business and the two point one million number landed somewhere familiar, sign up for Calabash Business at calabash.app/b2b. We will start with a real conversation, not a demo script.