Early AI returns are common. The capacity to scale them is not.
Three-quarters of enterprises report AI returns, but only a small fraction have the data, leadership and training in place to grow them.
Return on AI: Weekly Digest – August 4, 2026
Eight findings on deployment economics, workforce readiness and operational results, from warehouse robotics and supply chain agents to the true cost of owning AI.
Our take
The most important number this week is the distance between 76% and 6%. Most enterprises can point to AI returns somewhere, but very few have the data infrastructure to grow them, and only 3% of senior leaders say they are ready to guide adoption.
Two findings explain the gap. Wipfli shows that subscription fees are a small part of what AI actually costs once integration, maintenance and support are counted. And returns fall measurably when managers aren’t trained. The companies capturing durable value, like Walmart rerouting its supply chain ahead of storms, built the foundation first.
What to do about it
- Budget for the full cost of ownership, not just licenses.
- Train managers before, or alongside, their teams.
- Name who owns AI outcomes at the leadership level. Scaling needs an owner.