Credit Saison reports 954% ROI from ChatGPT Enterprise pilot

A pilot that beat its targets

Credit Saison's business centers on credit cards, with approximately 31.16 million cardholders, and it also runs financing and international businesses. Many of its employees work in generalist roles, and the company wanted to learn how AI could improve everyday work and produce tangible results.

Working with OpenAI, it ran a ChatGPT Enterprise pilot with 315 participants for a little over a month, from late June to early August 2025. The group included volunteers, employees nominated by the project team, every executive including the president, and representatives from every department. It made up just under 10% of the company's approximately 3,700 employees in Japan.

Credit Saison reports a 954% ROI and 14 hours saved per employee each month, equivalent to approximately 169 hours annually. Both figures significantly exceeded the initial targets of 500% ROI and 7.5 hours saved per person per month.

A post-pilot survey found that 75% of managers and 70% of nonmanagerial employees said they could not do without ChatGPT Enterprise. Approximately 80% of respondents in the legal department said it improved their ability to do their jobs, and more than 80% in the IT department reported greater job satisfaction.

Leaders and training drove adoption

The company's CoE, which leads AI adoption, used a dedicated Slack channel to share successes and failures openly, so individual experiences became lessons for the wider organization. OpenAI supported the effort with training for executives and practical workshops for beginners.

One highlight was a custom GPT built to simulate shareholder meetings. Employees could use voice mode to rehearse realistic, interactive exchanges. It drew strong interest from senior leadership, and requests to share the prompt followed, which helped spread voice-based AI across the company.

Kazutoshi Ono, Credit Saison's Senior Managing Executive Officer, CDO and CTO, pushed leaders to use the tool themselves: "It's not enough to encourage employees to use it. Leaders need to use it regularly themselves." Katsumi Mizuno, the representative director, executive president and COO, built his own custom GPT for company research and added it to his daily work. The company approved a companywide rollout within a month of the pilot's end.

Three use cases in sales

The source describes three examples of employees applying ChatGPT in their daily work.

• Retail sales: Staff at counters in shopping malls and other locations use a custom GPT for store performance reporting. It takes audio from the end-of-day debrief and summarizes each employee's results, meeting notes, handover items and focus areas. The company says it is now used at every location.
• Corporate sales: Teams use a custom GPT for company research to analyze customers, identify business challenges and assess where Credit Saison's services could help. In one analysis it surfaced challenges that employees had not fully recognized.
• Inside sales: A custom GPT for prospect roleplay plays the prospective customer, so less experienced employees can rehearse sales conversations by voice and get feedback afterward, without taking experienced colleagues away from selling.

From Employee AX to Corporate AX

Credit Saison describes AI transformation, or AX, in three dimensions. "Employee AX" changes how people work. In engineering, the company reports that Codex saves an average of 850 hours of work each month. "Customer AX" focuses on customer experiences, user interfaces and user experience, and the company is developing an AI call center in-house. "Corporate AX" rethinks organizational structures and roles around AI and is at an earlier stage, with specific applications still being identified. The company is also considering ChatGPT Work.

Ono said: "We can't stop at Employee AX. The challenge ahead is how far we can progress toward Corporate AX, redefining our organizational design and the division of responsibilities with AI as a starting point."

As AI takes on tasks autonomously, the company plans a more careful approach to security, with protection of customer information as a prerequisite. It intends to evaluate new capabilities and deploy them quickly once they are established as safe to use.

What it means

The case shows how a large consumer finance company moved from a limited pilot to a full rollout in about a month, based on its own measured targets. The figures are the company's own reports, published by OpenAI. Credit Saison says leaders using AI themselves and employees finding practical uses sustained adoption, and it plans to extend those improvements to customer interactions and the way the organization works.