Context and objectives

A leading logistics company faced a critical operational challenge: processing and verifying transport invoices.

This labor-intensive task required several full-time employees (FTEs) to handle thousands of emails annually. Transport companies worldwide sent invoices with critical attachments that needed careful verification to prevent overpayments for undelivered or incorrectly priced services.

Inaction carried substantial risks, including financial losses from missed deadlines penalties, payment errors, and overall administrative inefficiencies.

With Agilytic, the company implemented a streamlined process to:

• reduce staffing needs,

• strengthen quality control,

• and achieve significant cost savings.

The objective was straightforward: automating invoice processing to reduce administrative burden while ensuring accuracy and compliance.

Approach

We implemented a multi-phased strategy to address this challenge, starting with a thorough analysis of the existing workflow that revealed gaps in documentation and inefficiencies. We then went from Proof-of-Concept (POC) to full industrialization within a few weeks.

1. AI and LLM integration

We developed an invoice content extraction system using a multimodal Large Language Model (LLM). This system:

• converts invoices into HTML files while ensuring accurate data capture

• verifies company numbers on invoices against internal records to ensure compliance

• structures transport details and validates them against internal transport database using Python logic rules and Pandas to identify discrepancies

2. Administrative automation

To complete this automated invoice processing system, we partnered with B_Robots to implement UiPath. This allowed us to automate administrative tasks such as email attachment uploads, AI module triggers, and report updates.

3. Coaching and documentation

We provided comprehensive process documentation and coached the IT department to ensure a smooth transition to the industrialization phase.

Results

The project delivered exceptional results. The proof of concept demonstrated the potential to achieve annual labor cost savings of several thousand euros through implementation of the industrialized solution.

Beyond these cost savings, automating invoice processing freed up valuable personnel to focus on strategic tasks.

This project also served as the client's first successful AI implementation, creating momentum for innovation and sparking interest in exploring AI applications across the organization. The successful optimization of transport invoice processing has positioned the organization to extend similar automation strategies to other subsidiaries, driving broader operational efficiency and effectiveness.