07 / 07
Basic Withdrawal Tool
A small internal tool that records withdrawals in one fixed shape, with the fields that make a record auditable required rather than optional.
- TYPE
- Internal tool
- USE
- Withdrawal record keeping
- ROLE
- Design and build
- STATUS
- In use, ongoing
- CAPABILITIES
- Internal tooling, Operational systems

BEFORE
Records kept in whatever format each person preferred, individually fine, collectively unreconcilable.
AFTER
One shape, enforced on entry, so reconciliation is reading rather than reconstruction.
01 / INPUTWithdrawal records
02 / PROCESSOne format, enforced on entry
03 / DECISIONUniform over personal preference
04 / OUTPUTA trail that reconciles
Uniform
Records
Auditable
Trail
Deliberately small
Surface
The problem
Withdrawal records were being kept in whatever format the person keeping them preferred. Each format was individually fine. Collectively they were unreconcilable, and money records that can’t be reconciled are the ones that turn into an afternoon of archaeology the moment anyone asks a question about them.
What I built
A small internal tool that records withdrawals in one fixed shape, with the fields that make a record auditable (what, when, how much, by whom) required rather than optional.
How it works
The format is enforced at entry rather than corrected afterwards. A record that is missing one of those four fields cannot be saved, which is the only point at which enforcing it costs nothing.
Technical approach
There was an obvious temptation to grow this into approvals, roles, and a reporting layer. I didn’t, because every feature added to a tool that touches money is a new way for the record to be wrong, and the actual problem was never “we need more capability”. It was “we need the same capability, spelled the same way every time.”
A tool this small can be fully understood by the person using it. That property is worth more here than any feature I could have added.
Current state
In use. Withdrawal records are uniform and traceable, and reconciliation went from reconstruction to reading.