Enforce put its entire back office on one track

A business where the asset does not recover on its own

Enforce is a specialist manager of distressed and opportunistic assets, owned by a major investment bank. It runs four lines: corporate non-performing loans, judicial recovery and bankruptcy claims, distressed real estate, and public receivables. Founded in 2014, held in a joint venture with its parent bank from 2016, fully owned since 2021.

The model is to buy portfolios at a discount, clear the legal encumbrances, and recover value over a period of years. The company reports more than 300 employees across five offices, more than 10,000 properties cleared and sold, and more than 23,000 court cases tracked simultaneously across a network of partner law firms.

That last number shapes everything else. An asset in this business does not recover on its own. It recovers through thousands of small decisions, made by different people, under deadlines and approval limits, and frequently executed by third parties. Margin is not made at the purchase price of the portfolio. It is made in execution: what it costs and how long it takes to turn a stalled claim into cash.

So the back office here is not a support function. It is the production line. In an ordinary company a request lost in an inbox is an annoyance. In a distressed-asset manager it is one of three things: a vendor engaged without passing validation, a deadline nobody saw, or an approval sitting somewhere nobody can point to in an audit.

Four symptoms, one cause

The process existed. It just did not exist anywhere a system could see it.

“We had a lot of workflows here that were frankly a bit chaotic, because information had several ways in: sometimes HR received a headcount request by email, sometimes by chat, sometimes by phone. All of it stayed decentralized, when they didn't lose a request altogether.”

One detail changes how this case should be read. Enforce was already running a low-code platform before Jestor. This did not start at a company with no tooling. It started at a company that had already chosen to automate and needed a different cost and a better fit. The 50% comparison is between two platforms doing the same job, not between software and spreadsheets.

Item by item, what did each job before and what does it now

What was not replaced: the finance ERP. It stayed where it was, and Jestor now feeds it. That is deliberate. The tax and accounting system of a company this size is not where a project like this starts. What had no owner was the layer above it: who asks, who approves, by when, with which document. That layer was what lived in the nineteen spreadsheets.

The chain, from request to payment

1. 1
   Request form

   Open to the whole company, with required fields. The requester no longer decides who to send it to.

   Person
2. 2
   Routing by type and amount

   The automation reads the record and picks the approver. An administrative technology expense goes to the head of Technology, not to the requester's own head.

   Automated
3. 3
   Approvals in sequence

   Area head, director and the parent bank, depending on the amount.

   Automated
4. 4
   Document and signature

   Generated from the template. The signature stage is tracked on the record.

   External
5. 5
   Entry in the finance ERP

   A webhook writes the entry and attaches the payment receipt back to the record. No double data entry.

   Automated
6. 6
   Tracking dashboard

   Queue, deadline, approver and full history, per request.

   Automated

The chain used to break at the handoffs. Between the person who asked and the person who approves, between approval and contracting, between contracting and payment. Each of those boundaries was an email, and an email has no state: it cannot tell you whether it was read, whether it is stuck, or whose turn it is. Automating the handoff rather than the task is what closes the chain, and it is why the largest number on this page is not hours saved but notifications nobody had to write.

Every figure, with the basis beside it

What 26,000 notifications are worth. Enforce did not measure time saved, and this page will not invent a figure. But the order of magnitude is checkable. If each notification cost two minutes to write, address, attach and send, 26,000 of them come to roughly 870 hours in under a year. About five months of one full-time person, spent entirely on telling colleagues that something moved to the next stage. The two minutes are our assumption, not Enforce's measurement. At thirty seconds it still leaves more than 200 hours.

The hardest result to price is the 79 rejected vendors. Engaging a vendor without validation is not an administrative slip in this industry. It is legal, tax and reputational exposure that surfaces later, in an audit, when fixing it is already expensive. The gated track did not make the risk smaller. It moved the check in front of the signature.

“In a little under a year we've had almost 770 vendor review requests, with 677 registered. Vendors were sometimes engaged before that financial validation. In that period we had 79 rejected that we then could not engage.”

What this case does not measure

Worth naming, because it is usually what gets inflated.

Why cost per case is now the competition