AustraliaFinance and insurance2023 – ongoingWestpac
Generative AI coding assistants experiment
Westpac reported 46% higher productivity in an AI-assisted coding experiment
Problem
Westpac wanted to test whether generative AI helped engineers complete coding tasks without reducing code quality.
Approach
The bank randomly divided 60 engineers into 4 groups, including a control group that coded without AI. The other groups used tools from Microsoft, Amazon and OpenAI on the same assignments. Westpac checked the resulting code for vulnerabilities and maintainability.
Results
Takeaway
The result came from a short controlled experiment. It did not establish the productivity gain across everyday development work.
Categories
- Size
- Large enterprise
- Industry
- Finance and insurance
- Business function
- IT and R&D
- Innovation type
- Process innovation
- AI technology
- Generative AI
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Company's own claim
Sources
Source summaries
Westpac's CTO described the coding experiment and reported a 46% productivity gain. He said code-quality checks found no deterioration and initial teams had begun using AI through the bank's engineering platform.
Full details are in the original sources.
