A brief pilot with better-than-expected numbers
Credit Saison, a Japanese credit card operator, introduced ChatGPT Enterprise for all employees after a pilot conducted between late June and early August 2025, according to ITmedia AI+. The pilot involved 315 people, just under 10% of its roughly 3,700 domestic employees, including the president and the entire board of directors. The company had set targets of a 500% return on investment and 7.5 hours of work saved per person per month. The reported results were a 954% ROI and 14 hours saved per participant per month, equivalent to about 169 hours a year.
The decision to extend the tool across the entire organization came one month after the pilot ended. It is the figure that best sums up the scale of the operation: a very short verification phase, with better-than-expected numbers, followed by a rapid move to company-wide scale.
The role of senior leadership and sharing mistakes
At the basis of the adoption is a choice attributed to Kazutoshi Ono, a director and executive vice president with CDO and CTO responsibilities, according to whom it would not have been possible to understand the tool’s value without intensive and concrete use. Ono argued that top management should not merely invite employees to use it but should practice it themselves. President and COO Katsumi Mizuno reportedly developed a custom GPT for company research and uses it in his daily work.
The internal coordination group opened a dedicated Slack channel to share not only success stories but also failures, while OpenAI provided training for management and hands-on sessions for beginners. Among the uses that drew the most interest is a GPT for simulating the shareholders’ meeting, used as a dialogue exercise, with several requests to share the prompts.
Three operational applications and next steps
The material describes three concrete applications. The first is a GPT for customer service counters in shopping centers: starting from the audio recording of the end-of-day meeting, it automatically organizes results by staff member, the minutes, handovers for the following day, and the points to focus on, and it is used by employees at all stores. The second is a GPT for analyzing client companies, used by the corporate sales force to identify management problems and the contribution potential of the services offered; according to the source, applied to Credit Saison itself it reportedly brought to light issues not recognized internally. The third is a role-play GPT for inside sales, in which the AI plays a potential customer and enables voice practice with final feedback.
The company frames AI-driven transformation in three areas: Employee AX, Customer AX and Corporate AX. The first includes the reported reduction of 850 hours per month in the engineering department thanks to Codex; in the second, an AI-based call center is under internal development; the third is still being defined. The adoption of ChatGPT Work is also under study.
Some elements remain to be verified: the source does not detail how ROI and saved hours were calculated, nor the costs of the operation. The company also states that, precisely because AI can operate autonomously, it will pay particular attention to security and the protection of customer data, assessing results before extending new functions.
