Case studies

The week we handed invoicing to a machine.

The week we handed invoicing to a machine.

The client was a 14-person management consultancy. Their financial operations were handled by a part-time bookkeeper who came in twice a week, and an external accounting firm that handled year-end and tax.

The bookkeeper was good at her job. She had been with the firm for six years. She knew the suppliers, understood the quirks in the invoicing, and had a mental model of what the cashflow looked like on any given week. The problem was not her capability. The problem was what she spent her time on.

Two days a month, every month, went to one task: matching purchase invoices to purchase orders, checking for discrepancies, coding expenses to the right accounts, and chasing missing references. Roughly 16 hours of skilled, expensive time on work that is almost entirely mechanical.

What we built.

We deployed The Accountant with a specific initial scope: inbound invoice processing.

Every time a supplier invoice arrived by email, the system would read it, extract the key fields, match it against open purchase orders in their accounting software, post it if the match was clean, and flag it with a summary note if it was not. If an invoice had no matching purchase order at all, it would draft a query back to the supplier and hold it for human review.

Setup took three days. The first week it ran, the bookkeeper spent 40 minutes reviewing exceptions rather than processing everything herself.

What the numbers looked like.

In the first month:

  • 94 invoices processed
  • 81 posted automatically with no intervention
  • 13 flagged for review, all with a clear note explaining why
  • 0 mispostings that we caught in the month-end review

The bookkeeper's two invoice-processing days became half a morning. She spent the rest of the time on work that actually required her judgment: reconciling bank accounts, preparing management reports, handling the supplier relationships where something was genuinely wrong.

What the bookkeeper thought.

We asked her directly, because this matters. Her honest answer was that she had been slightly nervous before it started, and that she had spent the first two weeks checking every single posting manually to make sure it was right.

By week three she had stopped checking every one. By week four she trusted it on routine invoices and focused her review on exceptions.

"I thought I would feel like I was being replaced. I felt like I finally had an assistant."

That sentence is the best description of the experience we have heard.

What it changed beyond the obvious.

The firm's books are now current to within 48 hours, not the week-and-a-half lag they had before. That has knock-on effects. The owner can see actual cashflow numbers at any point rather than waiting for the bookkeeper's next visit. The finance manager they brought in two months later had live data to work from on day one.

The bookkeeper's rate did not change. Her hours dropped slightly, because she no longer needed to come in as often just for invoice processing. She used the time she freed up to take on a second client of her own.

The pattern.

This story is not unusual. We see a version of it in almost every firm we work with. Skilled people spending significant chunks of their time on work that is mechanical, repetitive, and frankly beneath what they were hired to do.

The work does not go away when you hire a digital employee. It gets done faster, more accurately, and without the bottleneck of a person who can only work two days a week.

The person moves up the value chain. The business moves faster. Both win.

Ready for a digital workforce?

One conversation. We design the right team for your business.

Book a call