In nearly every boardroom I sit in, the question is the same: "What's our AI strategy?" It is the wrong question, and BCG just put a number on why.
BCG's fourth annual AI at Work study, published in June 2026, surveyed 11,749 employees, managers, and leaders across 14 markets. They titled it "Strategy Matters More Than Tools." My read of their data is blunter: the companies pulling ahead are not the ones with the most licenses or the newest models. They are the ones who can tell you exactly which business outcome each deployment is meant to move.
Start with adoption, because it is no longer the story. In BCG's data, 74% of frontline workers are now regular AI users, a sharp jump from a year earlier. Adoption has gone mainstream. The tools are everywhere. What is scarce is direction, and BCG's numbers show what that costs.
Here is the finding I keep coming back to. BCG splits companies by what they actually do with AI: Deploy (bolt tools onto existing work), Reshape (redesign workflows end to end), and Invent (build new business models). The Deploy-only companies are stuck. The Reshape and Invent companies pull away on every measure that matters. Sixty-seven percent of their people see measurable business improvement, versus 43% at Deploy-only firms. Fifty-three percent save a full day a week, versus 31%. Same tools. Different results. The difference is not technology. It is whether the work was redesigned around a clear outcome.
Clarity is exactly where these companies diverge. At Reshape and Invent organizations, 52% of employees say their company's direction for AI is clear. At Deploy-only organizations, 31%. And those leading companies invest in their people at nearly twice the rate: 52% run a major reskilling effort, versus 27%. That gap matters, because the people problem is enormous. Eighty-eight percent of respondents say they will need significant upskilling in the next five years, and only 36% feel they have been trained for it. The tooling is solved. The readiness is not.
None of this surprised me. The case we keep making is that AI value is a people-and-clarity problem, not a technology one: roughly 70% of the return comes from how you redesign the work and prepare your people, and every initiative is really a choice between chasing efficiency and chasing growth. BCG's Deploy-versus-Reshape split is that same fork, measured at scale. The companies that treat AI as a strategy unto itself stall. The ones that put their existing strategy on AI compound.
So before you approve the next tool or the next pilot, make someone in the room finish this sentence: "This accelerates [specific outcome] by [specific mechanism]." If no one can, you have not found a technology gap. You have found a clarity gap. That is the one worth closing first.
Source: BCG, "AI at Work 2026: Strategy Matters More Than Tools," fourth edition, June 2026 (n=11,749 across 14 markets). Read the full report at bcg.com.