A month of bank reconciliation, done in 55 seconds.
Abdul Latif Jameel Health spent days every month manually matching bank transactions against Oracle. The ALJH AI Agent now reads the statement, matches by clear rules, flags exceptions, and delivers a CFO-ready report end to end in under a minute.
Days of manual matching, every single month
A large healthcare company in Saudi Arabia had a problem: every month, their finance team spent days manually matching every transaction on their bank statements against their Oracle accounting records. It was slow, repetitive, and stressful and the entire monthly close waited on it.
We built a system that does the same job in about one minute. It reads the bank statement automatically, matches transactions using clear rules any accountant would follow, flags the handful that need human attention, and produces a clean report the CFO can sign off onevery month, forever.
The client, Abdul Latif Jameel Health (ALJH), is part of the Abdul Latif Jameel Group a well-known name across the Middle East spanning automotive distribution, energy, health, and education.
Client Snapshot
Healthcare · Saudi Arabia
- Client
- Abdul Latif Jameel Health (ALJH)
- Parent Group
- Abdul Latif Jameel Group
- Industry
- Healthcare
- Region
- Saudi Arabia
- Accounting System
- Oracle
- Scope
- 3 Bank Accounts Main, Payroll, Supplier Payments
- Process Automated
- Monthly Bank Reconciliation
The Challenge
At the end of every month, the finance team had to confirm that every amount moving through three bank accounts main, payroll, and supplier payments was correctly recorded in Oracle. Each account carried dozens of transactions: customer payments, salary batches, supplier settlements, and bank fees.
Every transaction matched by hand, line by line, across spreadsheets and Oracle screens
Uncertainty on individual transactionswas this the right customer payment, the right salary batch?
Unexplained charges (e.g., small bank fees) requiring manual investigation
Internal transfers between accounts needing manual identification
The entire monthly close waiting on this single, manual step
Any small error meaning a fresh manual search to find it
This wasn't work anyone enjoyed and it wasn't work anyone should be doing by hand in 2026.
Eight structured steps, one minute end-to-end
Statement Received & Uploaded
The bank sends the monthly statement, which is uploaded into the client's Oracle system.
Currently manual automating this handoff is the next phase.
Automated Reading
The system reads the bank statement automatically.
Oracle Record Retrieval
It fetches every relevant record from the client's Oracle accounting system.
Rule-Based Matching
It matches each bank transaction to its Oracle counterpart using clear, accountant-friendly rules.
Exception Flagging
Anything that doesn't match is flagged, with a plain explanation of why.
Journal Entry Suggestions
For bank fees, the system suggests a journal entry to record them properly.
Balance Verification
Closing balances are checked to make sure everything adds up.
CFO-Ready Report
A clean, signable report is produced for the CFO with a full audit trail behind it.
Steps 3 through 8 reading the statement, matching against Oracle, flagging exceptions, and producing the final report run in about 55 seconds, end to end.
Numbers from a live production run
These figures come from an actual production run of the system in January 2026 not a demo or a projection. All numbers are real and can be substantiated on request.
Before vs After
Business Impact
By automating the most repetitive part of the monthly close, the finance team gets back days of work every single month time that now goes toward analysis and decision-making instead of line-by-line matching.
Because every match is backed by a clear rule and a confidence score, the CFO can sign off with confidence, and any exception comes with a plain-language explanation instead of a mystery to chase down.
The system is designed to be extended: today it covers three accounts and monthly reconciliation, with statement-upload automation planned as the next phase.
Why it's reliable
The agent is built to be trustworthy first and fast second. Core capabilities include:
Days of work.
Done in a minute. Every month.
Automate your monthly close.
If reconciliation, matching, or financial reporting is eating your team's time, we'll scope a production-ready agent rule-based, auditable, and CFO-ready.