Four out of five enterprise AI initiatives never reach the P&L. The difference between the ones that do and the ones that don’t has almost nothing to do with the model.
Walk into almost any mid-market or large enterprise today and you’ll find a wall of AI pilots. Walk back in a year and most of them are gone — quietly shelved, never scaled, rarely mourned. The technology worked. The value didn’t.
In our work across industrials, healthcare, and financial services, the pattern is consistent: the firms that capture real return from AI don’t start with the model. They start with a single operational decision or workflow where a better answer changes the outcome — and they tie the pilot to a specific P&L metric from day one.
The stalled pilots almost always share three traits. They are scoped around the technology rather than a decision. They have no owner accountable for the business result. And they have no path to production designed in from the start, so even a successful pilot dies in the gap between proof and scale.
The firms that win with AI don’t start with the model. They start with the decision the model is supposed to improve.
The 20% that succeed invert all three. They pick the decision first. They give it a business owner, not just a data-science owner. And they design the scaling path — data, integration, change — before the pilot begins, so success has somewhere to go.
For mid-market companies in particular, this is good news: you don’t need a moonshot. You need one high-value decision, instrumented properly, with someone accountable for the number it moves.
Key takeaways
- 01Scope AI around a decision, not a technology
- 02Give every pilot a business owner and a P&L metric
- 03Design the path to production before the pilot starts
Priya Nandakumar
Priya blends a consulting and operator background, having run a manufacturing operation before returning to advisory. She leads our digital, AI-in-operations, and supply-chain work — turning technology investment into measurable operational return.
