← Cases
AustraliaFinance and insurance2024 – ongoing

Bendigo and Adelaide Bank

Generative AI rewrite of the Agent Delivery System code

Bendigo and Adelaide Bank used generative AI to rewrite its teller app code in about three months

WorkingCompany's own claim

Problem

The Agent Delivery System is the teller software used by pharmacies and other non-bank agents. It ran on old code and needed modernising.

Approach

MongoDB's professional services team built generative AI tooling that worked with MongoDB Atlas. An internal large language model rewrote the code, so no data left the bank. Moving the database took about three weeks and the full modernisation about three months. Developers corrected some interpretations by hand. The bank planned the same approach for its branch teller system and payments platforms.

Results

about three monthsduration of the full modernisation, per the bank and MongoDBCompany's own claim[1]
90 percent of the way thereCIO's description of how complete the AI-generated code wasCompany's own claim[1]

Takeaway

The bank's CIO said the AI output was 90 percent of the way there, so human review stayed part of the work.

Categories

Business function
IT and R&D
Innovation type
Process innovation
AI technology
Generative AI
Evidence
Company's own claim

Sources

Source summaries

[1]iTnews13 Jun 2024

iTnews reported that the bank and MongoDB used generative AI to rewrite the Agent Delivery System. The database move took about three weeks and the whole project about three months. The CIO said the AI output was 90 percent of the way there.

Full details are in the original sources.

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