Report · The AI Productivity Plateau
91% feel faster. Only 12% can prove it.
A benchmark of AI adoption, productivity, and measurement from 427 responses from senior engineering leaders.
Key findings
- 01
91% of teams report higher productivity from AI, but only 12% are extremely confident in the metrics behind that claim — a 79-point gap between perceived and proven results.
- 02
Leaders expected more than they got: 49% believe effective adoption could produce gains of 51% or more, but only 23% report gains at that level today.
- 03
Investment keeps rising regardless. Even among teams that explicitly report a value plateau, 88% still plan to increase AI spending over the next six months.
Summary
Most AI rollouts began one engineer at a time: buy seats, encourage experimentation, and let teams find what works. That produced local speed, but it did not redesign review queues, testing, release processes, or the metrics used to judge the system. Adoption scaled faster than the operating model around it.
The result is the productivity plateau. 91% say AI made their teams more productive, yet only 45% report widespread team-level gains. Engineers experience local acceleration, but leaders cannot consistently connect it to delivery or business value.
Teams further along the adoption curve report stronger results: 71% of advanced teams report proven value, compared with 47% of teams at the widespread stage. Giving individual engineers better tools can increase local speed, but it cannot fix the bottlenecks that appear across review, testing, integration, and deployment. Individual gains become sustained delivery only when the system around them is redesigned.
What you’ll learn
Where the productivity plateau shows up, what gets in the way of scaling AI productivity, and a six-step working framework for measuring AI ROI — including a decision matrix for what to do when the numbers stall.
