An insourcing organization was set up from scratch in 2019. Currently, it has grown to nearly 200 people. Insourcing development started with the development of a smartphone app feature for credit cardholders. Today, it has expanded into a mission-critical system and it is promoting company-wide DX together with the information system division.
In terms of the streamlining of business, the software-based automation of manual work reached a cumulative total of 1.61 million work hours compared to 2019, equivalent to the automation of the work of nearly 800 employees. Furthermore, the amount of photocopying paper consumed has been reduced by a total of 102 tons. The cloud utilization rate for major systems has reached 80%.
We have introduced DX technologies to streamline businesses, and we have increased the speed and flexibility of new product and feature releases. Our next goal is the Credit Saison AI Transformation (CSAX), a drastic reform of business using AI.
1. Enabling all employees to be AI workers
What happens if all employees have the most advanced, highest-level AI? To answer that question, we distributed ChatGPT Enterprise to 315 employees and set a pilot period, from June to August 2025. Almost all of our departments, including sales, planning, call center, operation and system development, were involved in the demonstration experiment, which showed that the advanced AI could reduce work hours by 170 hours per employee (nearly 8.5% of annual work hours).
This led to a decision to distribute ChatGPT Enterprise to all employees. All employees will strive to evolve into AI workers who adeptly use AI and increase the sophistication of routine work.
2. Redesign of operations and AI-oriented business reform
Most operations in many operating companies, including ours, were designed before AI became what it is today. Going forward, we will dramatically reduce labor by redesigning operations assuming that AI tools are distributed to all employees and AI is a coworker.
The goal is to reduce work hours by a total of 3 million hours compared to 2019 by the spring of 2027.
3. AI-friendly information and system design
Conventionally, text, presentation materials and various regulations were created solely on the assumption that they will be read by humans. In the future, however, increasing the number of documents that AIs and humans are able to read will increase the likelihood that AI will be able to correctly understand in-house information and provide correct answers that are appropriate for a job.
Going forward, we will create and revise documents in an AI-friendly form and, whenever we design or develop a system, we will consider operability for humans and ease of access by AI.
4. Establishing AI governance
Undoubtedly, using AI increases convenience and productivity. However, it also entails risks, and some precautions must be kept in mind. For this reason, we will introduce the following seven measures to build a robust AI governance system.
1. Keep up to date on the status of the use of AI
2. Formulate AI principles and policy
3. Build an AI governance system
4. Formulate monitoring processes
5. Introduce monitoring tools
6. Develop human resources in the field of AI governance
7. Improve operating processes