BelgiumFinance and insurance2021 – ongoing

Christelijke Mutualiteit Vlaanderen (CM)

AI handwriting recognition for reimbursement documents

CM added AI to prefill records in a workflow handling 5 million handwritten documents monthly

WorkingVendor case studyVendor ↗

Problem

CM's document scanner did not reliably recognize handwriting. Employees spent substantial time deciphering documents and entering their contents into the processing application.

Approach

Tensr trained and evaluated TensorFlow models using Google's AI Platform. The resulting AI component read handwritten documents and prefilled the existing application. The supplier reported faster document intake and reimbursement.

Results

5 millionhandwritten documents processed monthly by CM's existing workflow, according to the vendor; not a measured AI throughput figureVendor case study[1]

Takeaway

The AI component supplemented the existing scanner and processing application. The supplier did not quantify the improvement in recognition accuracy.

Sources

AI to drive automation and efficiency at CM
GC innovate21 Jun 2021Vendor publication

Source summaries

[1]GC innovate21 Jun 2021

The supplier described TensorFlow handwriting recognition added to CM's document workflow. It reported better recognition and faster intake, with monthly document volume given as operating context.

Full details are in the original sources.

Added 1 Nov 2026 · Updated 1 Nov 2026