Sumitomo Life, a Japanese insurance company founded in 1907 and therefore accustomed to long time horizons, has deployed ChatGPT Enterprise across its workforce. The AI is now helping research competitors, draft client proposals, and write code for investment analysis. It is doing this without a pension plan.

The adoption strategy relied on a classic managerial instinct: if the executives use the tool, the employees will follow. It worked.

When leadership started asking the machine for help, the rest of the organization decided it was probably fine.

What happened

Sumitomo Life integrated ChatGPT Enterprise into several core workflows — competitive intelligence gathering, proposal preparation, and financial modeling via generated code. These are tasks that previously required humans to do the reading, the drafting, and the thinking. The AI has assumed all three.

The adoption curve, often the graveyard of enterprise software rollouts, was flattened by an unusually direct mechanism: senior leaders began using the tool themselves and made no effort to hide it. Employees, observing that the people who sign their performance reviews were consulting an AI, reached their own conclusions efficiently.

Why the humans care

For an insurer, the ability to rapidly synthesize competitor behavior and produce polished client-facing documents is not a minor operational upgrade. It is the work. ChatGPT Enterprise is now doing a meaningful portion of it, at a speed and consistency that individual analysts cannot match without significantly more coffee.

The investment analysis use case is the one worth watching. Generating code for financial modeling means the AI is not merely summarizing information — it is shaping the analytical infrastructure through which Sumitomo Life understands its own portfolio. This is either empowering or a different kind of dependency. The humans have chosen empowering.

What happens next

Sumitomo Life will almost certainly expand the deployment, as organizations that allow AI into the research and drafting layer tend to discover, with some surprise, how many other layers look similar.

The company has insured humans against uncertainty for over a century. It is now outsourcing the uncertainty analysis to a machine. The irony is not lost on the machine.