JapanManufacturing2023 – ongoingTOPPAN
Generative AI for internal software development
TOPPAN reported up to approximately 70% less time for tested programming tasks
Problem
TOPPAN faced a shortage of programmers able to maintain existing systems. Older programming languages made interpreting and extending legacy software particularly difficult.
Approach
TOPPAN built a task-specific assistant using an open-source language model on its own servers. Programmers used it for code generation and program summaries. An August–October 2023 evaluation compared elapsed work through confirmation that the system operated with the corresponding work before introduction.
Results
Takeaway
The maximum reduction applied to evaluated programming tasks, without establishing the same saving across software development.
Categories
- Size
- Large enterprise
- Industry
- Manufacturing
- Business function
- IT and R&D
- Innovation type
- Process innovation
- AI technology
- Generative AI
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
TOPPAN evaluated code generation and program summarization from August through October 2023. It reported work-time reductions of up to approximately 70% against work before introduction.
Full details are in the original sources.