Key takeaways
- The "AI multiplier" is real but unevenly distributed; companies with structured data and clean APIs see 30%+ productivity gains, others see negligible change.
- Knowledge-work productivity is the most-studied surface; the under-discussed one is operations and procurement.
- Sustained advantage requires re-architecting around AI as a primary user, not retrofitting existing systems with AI add-ons.
Two years into broad enterprise AI deployment, the productivity story is not “AI multiplies everything.” It’s “AI multiplies what was already structured.”
The multiplier is uneven
Companies with clean, accessible data and well-defined APIs see step-changes in throughput. Companies operating on a layer of legacy systems and undocumented processes see modest gains at best.
Operations: the underrated win
Most public AI productivity studies focus on knowledge-work — coding, content, research. The biggest wins we’ve seen are in operations: procurement automation, contract review, supplier onboarding. Less photogenic, more impactful.
Architecture matters more than tooling
The companies pulling ahead aren’t the ones with the most AI tools. They’re the ones who restructured their core systems to assume AI is a primary user — not a bolt-on.