Credit Saison: 954% ROI and Deployment of ChatGPT Enterprise

After a pilot lasting just over a month, Credit Saison reported a return on investment of 954% and gains of 14 hours per employee per month. The decision to roll out ChatGPT Enterprise company-wide followed within a month, with deployment to all employees scheduled for September 2025 and an AX strategy extending beyond internal uses.

Enhanced Security and Focus on Corporate AX Announced by Management

Credit Saison plans to further strengthen its security policy as AI takes on tasks autonomously, with the protection of customer information remaining a prerequisite. Kazutoshi Ono indicates that the next challenge is to move towards Corporate AX to rethink the organization and distribution of responsibilities with AI as the starting point. The company is structuring its transformation into three AX components, with Corporate AX in an early stage and ongoing identification of applications. It is also considering ChatGPT Work to focus employees on higher-value tasks. For customers, Credit Saison is developing an AI-based call center in-house. The company plans to evaluate new capabilities, improve its operations, and deploy quickly once security is established, viewing its relationship with OpenAI as essential to staying current.

Decision to Expand Within a Month After a Successful Pilot

The decision to extend ChatGPT Enterprise across the company was made one month after the pilot concluded, which had exceeded its time-saving objectives. In September 2025, the solution was deployed to all employees, giving everyone access to AI in their daily work.

A ROI of 954% and 14 Hours Saved Per Month Per Employee

The pilot produced a return on investment of 954% and average gains of 14 hours per month per employee, equating to approximately 169 hours per year. These results were observed in a group of 315 participants, forming the numerical basis for the scaling decision.

Governance, Training, and Leadership Engagement as Levers

The Center of Excellence facilitated a Slack channel to openly share successes and failures, fostering ownership. OpenAI provided training for leaders and workshops for beginners. A vocal GPT simulating shareholder meetings generated significant interest from management and quick requests for prompt sharing, promoting the adoption of voice AI. At the initiative of Kazutoshi Ono, Katsumi Mizuno utilized AI on the front lines, creating a business research GPT integrated into his daily routine. Adoption was supported by user leaders and employees applying AI to their tasks.

Operational Uses in Sales and Team Satisfaction

In retail sales, a reporting GPT helps document end-of-day meetings. In corporate sales, a research GPT is used to analyze clients, identify challenges, and assess service fit. In internal sales, a role-playing GPT provides additional training for new hires without burdening experts. According to a post-pilot survey, 75% of managers and 70% of non-managers reported they could not do without the tool. About 80% of legal respondents perceive an improvement in their work capacity, and over 80% of IT respondents report increased satisfaction.

Scope of the Pilot, Participant Profile, and Company Context

The pilot brought together 315 participants from late June to early August 2025, consisting of volunteers, designated employees, all executives including the president, and representatives from each department. They represented just under 10% of approximately 3,700 employees in Japan, in a company where many positions are generalist. Credit Saison, a financial player primarily focused on credit cards, also reports saving an average of 850 hours per month through Codex within engineering. The company operates in the Asia-Pacific and Oceania regions and is part of the financial sector.