The first wave of generative AI produced impressive demonstrations and a long list of pilots. The second wave is about a harder question: what has changed in how the business performs?
Start from the decision
Pilots that scale tend to share one trait. They target a specific, frequent decision or task with a measurable cost: pricing a quote, triaging a claim, drafting a credit memo, routing a delivery. The value case is written before the first prompt.
What separates scaled programmes
- Ownership in the business. A business leader owns the outcome. Technology teams own the platform.
- Data that is good enough. Not perfect, but governed, accessible and trusted for the use case at hand.
- Redesigned work. The tool changes the process, roles and incentives around it, or adoption stalls.
- A portfolio view. A small number of high-value use cases funded properly, rather than dozens of experiments funded thinly.
Where to begin
Pick three decisions that matter, measure how they are made today, and redesign them with AI in the loop. The results will tell you where to go next.
