← Cases
ChinaFinance and insurance2026 – ongoing

Bank of Shanghai

LLM parsing of custody fund-transaction files

Bank of Shanghai reported over 80% higher batch-processing efficiency for fund files

WorkingCompany's own claim

Problem

Custody staff handled fund files with inconsistent fields and layouts.

Approach

AI engineers and business staff introduced LLM parsing. The workflow produced standardized output.

Results

80%+batch-processing efficiency improvement over manual work, per Bank of ShanghaiCompany's own claim[1]

Takeaway

The application addressed file-format variation in custody operations.

Categories

Business function
Production and quality
Innovation type
Process innovation
AI technology
Generative AI
Evidence
Company's own claim

Sources

  1. 上海银行以AI实践和数字化转型赋能创新发展
    Shanghai SASAC21 Jul 2026Government publication

Source summaries

[1]Shanghai SASAC21 Jul 2026

Shanghai SASAC described operational LLM file parsing. It reported higher batch-processing efficiency than manual work.

Full details are in the original sources.

Added 1 Nov 2026 · Updated 1 Nov 2026
Open case page ↗