Our new report, A Judgement Layer for GenAI, sets out a better starting point for regulated businesses faced with accelerating their work with AI. In short: work out where your people’s judgement creates value before you choose a single tool or use case.
Get the full report, A Judgement Layer for GenAI
Avoid the wrong first move
An impulsive move would be to hand GenAI to the IT team, commission a list of use cases and wait for the value to arrive. But technology can only answer the ‘how’; it never asks whether the work is worth doing in the first place. Without a deliberate method, you end up tidying the edges of the business when the real work was needed in the middle.
GenAI raises the stakes. Cloud and SaaS changed how work gets done. GenAI changes what gets done, and that forces two questions at once: which work still earns its right to exist, and how your operating model and workforce change on Monday morning once you’ve automated it. Miss either and the initiative fails.
Judgement first, then tools
In advisory, underwriting, pensions and wealth management, clients pay a premium for hard-won expertise. They don’t pay for the hours experts lose chasing sign-offs, digging numbers out of emails and making the fourth round of ‘final’ tweaks to a document.
That makes GenAI a design problem about the shape of the work, and the responsibility sits with business leadership rather than the technology function.
The paper sets out three starting points to run in parallel: setting the direction of travel, knowing your worth through Outcome-Based Service Design, and building capability by getting the tools into the hands of your experts.
It also makes the case for being picky about what you automate. Deciding where human expertise stays is a design decision you make case by case, and the paper names the four reasons that justify keeping a human in the loop. Anything else is likely to be resistance dressed up as caution.
In the full report, you’ll find the three starting points in detail, the four tests for keeping a human in the loop, and how a specialist insurer moved from better tooling to better judgement, on target to reclaim 200 hours a week of its expert underwriters’ time. It closes with three questions to test where your own organisation should start.
