JapanFinance and insurance2025 – ongoingMizuho Bank
AI-OCR digitization of trade documents
Mizuho Bank cut processing load and lead time on trade documents by about 50% with AI-OCR
Problem
Staff in Mizuho Bank's international trade operations checked company, representative and country names on paper trade documents and keyed them into several systems. A single transaction could run to several thousand pages, and hundreds were processed daily. Tighter sanctions and financial-crime monitoring made accurate name entry more important.
Approach
IBM Consulting and IBM Technology Expert Labs built an AI-OCR workflow with the bank. A scanner turns mailed documents into PDFs, IBM Datacap extracts and classifies the information, and Watson Discovery structures the text and corrects it against a dictionary. An operator checks the result, and the data flows to downstream systems without double entry. The bank plans to extend the approach to other paper-based departments.
Results
Takeaway
Document AI paid off where staff retyped the same data into several systems. A human check stayed in the loop for accuracy.
Categories
- Size
- Large enterprise
- Industry
- Finance and insurance
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Partner co-development
- Evidence
- Vendor case study
Sources
Source summaries
IBM describes an AI-OCR workflow for trade documents in a Mizuho Bank operations department, using Datacap and Watson Discovery. The page reports a reduction of about 50% in processing load and lead time. It says the bank plans to extend the approach to other departments.
Full details are in the original sources.
