
In private conversations, most engineering leaders admit the same thing: their public AI narrative is running ahead of what their teams have actually shipped. New research puts a number on it — roughly four in five feel pressure to overstate progress to boards and investors.
Where the Pressure Comes From
Board-level AI mandates rarely come with a realistic timeline attached. Leaders are asked to show momentum quarter over quarter, even when the underlying integration work — data pipelines, evaluation harnesses, security review — takes longer than a slide deck implies.
A more honest way to report AI progress
- Report adoption metrics, not just pilot counts
- Separate "in production" from "in evaluation" explicitly
- Give engineering a seat in setting the external narrative
The gap between the narrative and the reality is exactly where AI initiatives quietly stall — reported as wins internally, never actually adopted by the teams meant to use them.


