Bank statements from 30+ banks, read and checked line by line.
A financial-services group was reviewing bank statements by hand. CONE RED built the system that now reads each statement into one checked database, so their people review exceptions instead of retyping. In daily use since autumn 2025.
As of June 2026, in production
The challenge
Bank statements arrive in dozens of layouts: every bank formats dates, running balances, and debit/credit columns differently. Reviewing them by hand is slow, and a single mis-read of a transaction’s direction (money in vs. money out) corrupts every downstream reconciliation. The client needed statements turned into clean, structured, correctly-signed transaction data without a human checking every line.
What we built
CONE RED built a system that reads each bank’s own statement format and checks it line by line into one database. Accountants load the statements, the system reads them, and a manager reviews the result in reports. Accuracy is measured from the corrections its users make, so quality is a number the client can see rather than a claim they have to trust. In December 2025 the system had to move to a new server; we moved it that month and every record came across.
The measured result
As of June 2026 the system had read about 1,600 statements and about 11,400 transactions from more than 30 banks and payment providers, with about 99% of transaction amounts right, measured from the corrections its users made. Those are figures from live use, not a projection.
A note on these numbers. The figures above come from one client’s production use, as of June 2026. They are specific to that client’s documents and volumes; results vary by engagement, document quality and scope. AI systems are probabilistic, so we report measured accuracy rather than promising a fixed figure. Client identity withheld.
Have a document-processing bottleneck?
If your team is re-keying data from PDFs, statements, or forms, we can scope an automation that reports its own accuracy.
Talk to us