Handwritten claims and automating backend workflows across 40+ clinics

A multi-specialty clinic network in Saudi Arabia replaced manual claim entry with Sarj document agents. Handwritten claim forms from more than 40 clinics are now read, checked and filed automatically, with 80% less manual entry and claims submitted within a day of the visit.

The network's doctors still write on paper at the point of care. That suits the consultation, but every form then had to be typed into the claims system by hand before it could go to an insurer.

The challenge

Each clinic sent its paper forms to a central back office, where a team re-keyed them: patient details, diagnosis and procedure codes, and the doctor's notes. Handwriting varied from clinic to clinic, forms mixed Arabic and English, and errors were found only when a payer rejected the claim.

The backlog grew with every new clinic. Claims routinely waited days to be filed, which slowed the network's cash flow and left little time to fix rejections.

The network also needed a solution it could run on infrastructure it trusted. Patient records could not be sent to services outside the Kingdom.

Why Sarj

• Arabic and English handwriting. Sarj's document models read mixed-language, handwritten forms, including the abbreviations clinicians actually use.
• Workflow, not just extraction. Sarj agents check what they read against the network's rules and file the claim in its systems, rather than handing back a text file.
• In-Kingdom hosting. All documents and data are processed on sovereign cloud infrastructure in Saudi Arabia.

The solution

1. Capture at the clinic

Front-desk staff scan each form when the visit ends. There is no new device and no change for the doctor.

2. Read and check

Sarj reads the handwritten fields, maps diagnoses and procedures to the right codes, and checks the claim against each payer's requirements. Anything it is not confident about goes to a reviewer with the original form beside the extracted fields.

3. File and track

Approved claims are filed in the network's claims system automatically, and each one is tracked until the payer responds. Rejections come back to the same queue with the reason attached.

Results

• 40+ clinics feed one claims pipeline.
• 80% less manual claim entry for the back-office team.
• Under 24 hours from the patient's visit to claim submission, down from several days.
• Every claim traceable from the scanned form to the payer's response.

Conclusion

The network kept the paper forms its doctors prefer and removed the typing that came after them. Claims move faster, the back office scales with the network, and every document stays in the Kingdom.

The network is now applying the same agents to referral letters and to insurance pre-approvals.