The AI Gap is Expanding
Lately, discussions about AI-driven changes in the corporate world often resemble homeowner debates about remodeling: a upgrade would add real value, let’s do it this year, we’ll proceed once plans are ready and the budget is set. Then progress stalls. Weeks turn into months, and before you know it, a year has passed.
Meanwhile, costs rise and new issues emerge—issues that the very upgrade could have helped prevent. The harsh truth sets in: you may think you can’t afford to change, but you can’t afford not to.
BCG’s Build for the Future 2025 report reveals a striking reality: only about 5 percent of companies are extracting substantial value from AI. That small group is achieving roughly twice the revenue gains and 40 percent greater cost reductions compared with peers who haven’t invested in AI capabilities. These 5 percent have learned how to generate real, scalable value from AI—and their advantage is likely to grow as technology progresses. In contrast, more than 1,200 companies studied show that 60 percent have little to show for their investments. The gap between early adopters and slow movers isn’t a narrow margin anymore—it’s a widening chasm. The reason? many firms mistake cautiousness for deliberate stagnation.
“Leaders are pulling away from the pack more and more,” says Nicolas de Bellefonds, global AI leader at BCG and co-author of the report. “The era of passive watching is over.” This isn’t just guidance to adopt AI—it’s a finding from the study itself. AI technology is advancing so rapidly that every week spent observing rather than acting makes catching up increasingly difficult. Future-built companies—BCG’s term for the AI leaders—are “pulling away, widening the value gap, and sinking slow movers deeper into a value hole.”
To uncover how these future-built firms achieved their status, the report analyzes how they deploy AI and allocate their spending. It finds that the biggest gains come not from isolated pilots or sporadic chatbots, but from transforming core areas—sales, manufacturing, supply chain, R&D, and increasingly IT—across the entire organization. “Major value comes not from AI pilots or standalone use cases,” the report notes, “but from reengineering core business workflows end-to-end.”
Amanda Luther, Global Leader of AI and Digital Transformation at BCG and a co-author, emphasizes that “focusing on a few bold bets” that ripple across the value chain can deliver the strongest ROI. “If you’re in consumer goods, target marketing or R&D for a material lift. In industrial sectors, prioritize manufacturing and supply chain,” she explains. “The goal is to identify those handful of areas that truly matter for your competitive edge and drive hard there.”
The report isn’t merely a retrospective on how a select few became AI leaders; it’s a blueprint for fast followers. It highlights that true progress isn’t about catching up with every possible use case, but about moving now and collecting data to prove outcomes.
“For those who haven’t moved yet, there’s good news: there’s proven, scalable value at other organizations,” Luther notes. “You can become a fast follower by concentrating on the areas where real value exists. Yet the gap is widening. Early adopters have built strong data foundations and an organizational muscle for adopting technology and change that will keep getting stronger.”
De Bellefonds agrees, cautioning against measuring success by the sheer number of use cases. Leaders are recognizing AI’s value and reinvesting returns, with study participants planning to double their IT spending this year and direct a larger share of that budget toward AI.
Agentic AI—the kind of AI that reasons, learns, and acts autonomously across workflows—stands out as one of the fastest accelerators. Once deemed a buzzword, it now accounts for roughly 17 percent of AI value in 2025 and could rise to about 29 percent by 2028. Deep integration of agentic AI into strategically important areas—such as marketing or supply chain—appears to be best practice for realizing that value.
As with any high-ROI transformation, progress requires time and effort.
“There’s a common myth that AI is a magic wand that makes everything easy,” says de Bellefonds. “That misconception is often promoted by some AI vendors. Real transformation demands hard work—changing how you operate, not merely deploying a tool. If you skip the foundational work, including operating models, skills, and guardrails, you’ll struggle to scale, no matter how sleek your prompts or clean your code.”
The study suggests that AI failures are rarely about the model itself. Most obstacles are human and organizational—misaligned strategies, training gaps, insufficient adoption, unstructured data, and vague goals. Firms that neglect foundational pieces fail to scale, regardless of how polished their prompts or code might be.
Luther shares a practical rule of thumb she and her team used: 10 percent of AI success comes from the algorithms, 20 percent from the technology, and 70 percent from the people. The longitudinal findings confirm that the people factor is decisive. When done well, AI reduces tedious tasks and liberates people to focus on judgment and creativity.
The bottom line is: the business case for AI is also a human case, demanding thoughtful people-centered execution as much as technical prowess.
.jpg)
