Employees at Sumitomo Life who used ChatGPT more than 100 times a month reported an estimated 524% return on investment, OpenAI reported on November 18.
Infrequent users, by contrast, logged an estimated 55% — nearly a tenfold gap inside one company.
That spread, not the headline number, is the real story in the Japanese life insurer’s deployment. The overall estimated ROI across the company averages 203%, based on employee-reported time savings, and weekly usage of ChatGPT Enterprise has passed 80%. But the figures move sharply with frequency: 55% at 8 to 15 uses per month, 231% at 16 to 30, and 524% at 101 or more. Adoption intensity, not license count, is what the insurer is actually measuring.
OpenAI’s case study, published Tuesday, credits the company’s leadership for that intensity. Executives and managers were among the first users, starting with a pilot of ChatGPT Business in 2025 involving roughly 100 to 150 people, including planning teams. The president used AI himself during the pilot and repeatedly encouraged employees to do the same in management meetings and on the internal portal.
“When senior leaders don’t understand a new technology, it can be hard to get it off the ground. Here, our executives and managers were among the first to use it. They saw for themselves what it could do. I think that made it easier for everyone else to start using it.”
said Kakeru Tateyama, a senior member of Sumitomo Life’s planning department, in the case study OpenAI published.

From chat queries to Enterprise
The rollout was slow by design. Sumitomo Life began using generative AI at work in 2023, mostly as a question-answering chat tool. ChatGPT stood out among the general-purpose tools the company tested for its range — deep research across sources, plus integration into work already happening in Word, PowerPoint, and Excel. In June 2026, the insurer moved from the Business pilot to ChatGPT Enterprise.
The security review mattered as much as the features. As a financial institution, Sumitomo Life worked with OpenAI to assess the Enterprise tier against its own standards, covering workspace access controls, feature permissions, and OpenAI’s policy of excluding business data from model training by default.
Where the 524% comes from
The heaviest users are not casual chatters. Planning teams use ChatGPT to track developments at other Japanese insurers and find overseas examples, then combine that research with firsthand information to build proposals for senior management. Investment analysts use it to draft code for statistical analysis of returns and risks such as price swings. Accounting and finance teams use it for research and analysis.
Projects in ChatGPT keep that material organized — related chats, files, and custom instructions in one place. That workflow persistence is likely part of what separates the 101-plus cohort from everyone else: the tool stops being a search box and becomes infrastructure.
| Monthly uses | Estimated ROI |
|---|---|
| 8 to 15 | 55% |
| 16 to 30 | 231% |
| 101 or more | 524% |
| Company overall | 203% |
OpenAI notes the estimates rest on employee-reported time savings — self-reported data, not audited productivity telemetry. The insurer also looks beyond hours saved in its quarterly survey, asking how often employees use ChatGPT, for which steps of their work, and whether they use it with colleagues.
Employees are beginning to take on work they previously lacked the time or specialist skills to attempt, a benefit the company says is harder to capture in a number.
Redrawing the work itself
The most pointed observation in the case study comes not from OpenAI but from inside the planning department. Masakazu Obi, a deputy general manager there, argues that bolted-on AI leaves most of the value on the table.
“If we take our existing ways of working for granted, we won’t get the full benefit of AI. The question is whether we can change the processes themselves. As we introduce AI, we’re also working to change those processes over the long term.”
said Obi, in the OpenAI report.
That is the harder version of the deployment — process redesign rather than tool adoption — and it is what the insurer says comes next. Expansion to more employees begins in the second half of fiscal 2026, with Sumitomo Life identifying, role by role, which tasks AI can handle, where humans work alongside it, and which responsibilities stay with people.
Over the longer term, the company is considering AI for more specialized work including underwriting and claims payments, and plans to explore new technologies with OpenAI as it redesigns workflows.
Time back for the customer
The stated end goal is not headcount reduction. Sumitomo Life frames the whole program as buying back time for conversations with customers, which it ties to the well-being of both customers and employees. Kuniaki Nakagawa, manager of the company’s AI Strategy Task Force, put it this way in the case study:
“For us, the value of insurance is the peace of mind it brings to our customers and to society. There’s something only people can offer, and it comes from talking with our customers and understanding how they feel.”
A three-year arc — from chat experiments in 2023, to a 100-to-150-person pilot in 2025, to Enterprise in June 2026 — ending with a claim about human conversation is an unusual close for an AI case study. Whether the 524% holds as usage spreads to thousands of employees, and whether process redesign outpaces tool rollout, is what the second half of fiscal 2026 will show.