---
title: "Customer case: CIMS Netherlands"
description: An invoice processing system with OCR, AI and score-based matching for a licensed Port Reception Facility with 25+ years' experience. Delivered in six weeks.
url: "https://rual.nl/en/customer-cases/cims-netherlands"
locale: en
alternate: "https://rual.nl/customer-cases/cims-netherlands"
updated: 2026-08-03
source: "https://rual.nl"
---

Maritime · waste processing

# CIMS Netherlands

Purchase invoices pulled from the mailbox, read with OCR and AI, and matched automatically to the right order lines.

![Two CIMS crew members connecting a hose on the deck of a tanker](https://rual.nl/site-assets/_opt/customers/cims-people-960.webp)

- **25+ years** in complex waste processing

- **6 weeks** from start to delivery

- **Q4 2025** delivered

## The situation

CIMS is a fully licensed Port Reception Facility with more than 25 years of experience in complex waste processing for the maritime, offshore and oil and gas sectors. Every job in the port produces paperwork, and that paperwork arrives from dozens of suppliers at once.

Purchase invoices landed in the mailbox and were laid next to the matching order by hand: line by line, comparing date, price, description and job number. Slow work that goes wrong easily and adds nothing.

## What we built

An invoice processing system that pulls purchase invoices out of the mailbox automatically and reads them with OCR and AI. It then matches the invoice lines to the existing order lines on several characteristics at once.

At its centre is a matching algorithm that scores on date, price, description, job number and other relationships. What is certain goes straight through. What is doubtful appears in a review screen where an employee can check, correct and approve it at a glance.

From the same environment, approved invoices flow through to the accounts in Elvy, with a dashboard on top showing what is coming in, what is stuck and why.

## The hard part: matching without trusting blindly

An invoice matched to the wrong order automatically is worse than an invoice left waiting. So the system is built around scores rather than a hard yes or no: every match carries its reasoning, and only above a confidence threshold does it pass through untouched.

The system learns from earlier decisions, so its suggestions improve as more is processed. That has cut the manual cross-checking back sharply, without anyone giving up control.

### Connections in this project

- Elvy accounting
- Mailbox connection
- OCR processing
- AI recognition

### What it delivers

- Invoices processed straight out of the mailbox
- Invoice lines matched on scores
- Approval workflow in the same environment
- Flow-through to the accounts in Elvy
- A dashboard with the status of every invoice
- Recognition that improves as more is processed

| Term | Def |
| --- | --- |
| Sector | Maritime waste processing |
| Build period | Q3 to Q4 2025, delivered within six weeks |
| Team | 1 part-time developer and 1 part-time architect, alongside the client |

![The deck of a tanker with pipework, alongside a large sea-going vessel](https://rual.nl/site-assets/_opt/customers/cims-ship-960.webp)

Alongside in port: where the stream of invoices comes from

## What happened next

The system was delivered within six weeks and is being extended on the same platform, from better recognition to additional reporting.

## A back office that largely keeps itself up to date.

Processing now runs automatically, with human review on the exceptions. The multi-tenant setup means the same environment can serve other companies in the group as well.

## More customer cases

Other operations running on our platform every day.

[All customer cases →](https://rual.nl/en/customer-cases)

- **[W.K. Heyl Jr.](https://rual.nl/en/customer-cases/heyl)** — Logistics · warehouse automation

![The W.K. Heyl Jr. distribution centre with green trucks at the loading docks](https://rual.nl/site-assets/_opt/customers/heyl-960.webp)

- **[Holstein Flowers](https://rual.nl/en/customer-cases/holstein-flowers)** — Floriculture · order processing & auction integration

![The Holstein Flowers premises](https://rual.nl/site-assets/_opt/customers/holstein-flowers-960.webp)

- **[Centercom Buitenreclame](https://rual.nl/en/customer-cases/centercom)** — Media · planning & placement

![A0 advertising frames on a city building, with traffic passing in front](https://rual.nl/site-assets/_opt/customers/centercom-a0-frames-960.webp)

## Is your back office full of retyping too?

Six weeks later most of it can run automatically, with review kept on the exceptions.

[Book a call](https://rual.nl/en/demo)

[All customer cases](https://rual.nl/en/customer-cases)