OpenAI’s launch of ChatGPT in 2022 was the starting gun for a once-in-a-generation disruption that is only beginning to reshape how the world works. Yet despite the transformative potential of AI to redefine every industry, most companies remain stuck in experimentation mode. Quarter after quarter, our survey of enterprises tells the same story: Fewer than 20% have scaled their generative AI efforts in any meaningful way.

The business world is stuck.

We know many of the reasons for this gridlock: technology deployed without a clear tie to business outcomes, pilots lacking focus or specificity, and an overwhelming sprawl of use cases with no prioritization across the vast AI landscape. But at the heart of the matter lies a problem that’s easy to diagnose and hard to fix: Most organizations are treating generative AI as a technology deployment rather than a business transformation.

Unlike previous technology waves, gen AI doesn’t create value through basic adoption. ROI comes from reimagining how work gets done and how a company competes. And that requires something deeper: business redesign with AI at the core.

Instead, many companies are falling into what we call the “micro-productivity trap”—a proliferation of proofs of concept and isolated use cases that deliver modest, localized efficiency gains but fail to scale. Tools get deployed. Demos impress. But outcomes never materialize.

This trap isn’t just a missed opportunity; it’s a growing strategic risk. Leadership teams see it themselves: AI battlegrounds are emerging across every sector. And as gen AI evolves from text to multimodal, from passive to physical, from tool to autonomous agent, the risks of delay compounds. Competitive advantage is at stake.

For companies already deep in experimentation, the challenge isn’t starting; it’s restarting. Resetting means stepping back to ask: Are we focused on outcomes or activities? Are we spreading ourselves too thin? Are we building the muscle to scale? The emerging leaders that are finding their way out of the trap are recommitting—top-down—to four moves that consistently separate AI leaders from laggards.

Top-down leadership

It sounds obvious, but in our experience, this is the critical starting point that many companies and boards haven’t fully committed to: Leadership matters. AI transformation starts and succeeds in the C-suite.

Grassroots experimentation with AI tools sparks innovation and cultural momentum—all good—but it does not self-organize into enterprise-wide impact. Without clear direction from the top, these efforts remain fragmented, siloed, and ultimately shallow. Why? Because teams need permission and protection from their leaders to both take on and adopt change of this magnitude.

The companies making real progress integrate gen AI into the heart of their strategy, with ambition matched by executive ownership. They set bold ambitions grounded in business strategy. They articulate a clear point of view on how AI will reshape their industry and how they intend to lead, not follow. And they set the tone from the top.

One telling example is Shopify CEO Tobi Lütke. In April 2025, he issued an internal directive mandating that all Shopify employees integrate AI into their daily work. He declared AI usage a “baseline expectation,” requiring teams to demonstrate why AI couldn’t perform a task before requesting more resources or headcount.

We’ve seen other companies take similar steps to make AI transformation a strategic and cultural priority. Some are rewriting incentive systems and embedding AI-related objectives into performance reviews, bonus structures, and promotion criteria. Others are launching large-scale upskilling initiatives—courses, competitions, and microlearning programs—to build AI fluency across functions.

But the baseline is clear: In the companies making real progress, executive teams are hands-on—setting the agenda, reinforcing the vision, using AI themselves, and making AI adoption a clear and visible priority. Sponsorship isn’t performative; it’s active, intentional, and tied to outcomes.

Fewer, bigger bets

Ambition is necessary but not sufficient. AI opens up a thousand possibilities. The most successful companies make focused, grounded bets and resist the urge to spread AI everywhere without real and specific outcomes in mind. They avoid the “thousand points of light” approach that has defined the early months of gen AI deployments: dozens or even hundreds of disconnected pilots that never scale.

Instead, they identify four to five critical domains—clusters of high-impact, interrelated use cases—and concentrate their transformation efforts there.

Navigating AI disruption this way is both strategically smart and financially material. Across industries, we see a handful of domains emerging as battlegrounds, places where competitive advantage will be won or lost. In technology, it's the software development lifecycle. In healthcare, it's drug discovery, regulatory management, and patient engagement. In retail and consumer products, it’s areas such as personalization, content creation, dynamic pricing, and demand forecasting. These aren’t just opportunities; they're the front lines of industry transformation.

And crucially, these domains aren’t standalone use cases; they’re systems of work. The software development lifecycle, for example, includes more than 40 discrete use cases. With less than half of developer time spent “hands on keyboard,” copilots alone are insufficient. Meaningful productivity gains require coordinated changes across design, testing, code review, and planning.

B2B sales is another example: One use case rarely delivers real impact, because go-to-market work is fragmented across dozens of micro-tasks. Unlocking more customer-facing time for commercial teams and higher conversion requires systematically rewiring the entire sales lifecycle, from lead generation to quoting and closing.

Transformation happens when organizations think in terms of domains that drive competitive advantage and real ROI, not point solutions. The companies that succeed don’t guess. They do the hard work upfront: defining the right domains, setting top-down value hypotheses, and building the mechanisms to measure, manage, and scale reinvention over time.

Process redesign from the ground up

You can’t automate your way to transformation. You have to rethink the work itself.

True gen AI impact requires detailed, zero-based process design: mapping where you are today (the “point of departure”) and reimagining how the work could operate with AI embedded from the ground up (the “point of arrival”).

This isn’t about layering tools onto broken workflows. It’s about building entirely new processes with gen AI at the center. And in our experience, it’s the process redesign—not the technology—that creates most of the value.

One compelling example is the way a major bank transformed customer engagement with AI. After years of disciplined investment to build a digital foundation that provided a 360-degree view of customers, the bank had ample evidence of the value of using that intelligence to engage customers: In selected campaigns, customer lifetime value doubled, and customer advocacy increased threefold, measured by Net Promoter System^®. But scaling that approach to all 18 million of its customers was a monumental challenge.

A key step was redesigning the process with AI-native workflows. The bank created a new business area with a small number of teams focused on “customer missions,” improving engagement at key moments in the customer lifecycle. Rather than use campaigns to push sales messages at specific points in time, these teams use intelligent “triggers” to engage customers with context and helpful options. For example, the bank sees that a customer has been withdrawing cash from an ATM that charges a fee and sends them a notification highlighting ATMs in their area that don’t.

The bank’s strong tech foundation combined with agentic AI technology—which enables models to do complex reasoning and problem solving on the fly—meant that a huge volume of manual work could be automated, freeing up the customer mission teams to focus on generating data-driven ideas with high impact. The teams use a purpose-built AI tool that allows them to understand customer insights, model different engagement strategies, quickly transform the best ideas into tests, and then continuously measure and optimize customer engagement based on what’s working and what’s not.

The results so far? Turning a customer insight into a campaign that’s in-market now takes one day, compared with 60 to 100 days previously. What once required 40 employees and 10 handoffs is now accomplished by four or five employees with no handoffs.

This work is often not glamorous; it requires understanding today’s workflows in detail and having the imagination and commitment to rebuild them from the ground up. But it’s what separates marginal gains from step-change performance.

An operating model that delivers transformation

Everything described above is hard: strategic prioritization, setting specific future-back goals, detailed process mapping, smart tech deployment, behavior change, and governance. While outcomes, not tools, must remain at the heart of the transformation, the underlying technology, data, and security foundations are equally vital. Getting those right requires rigorous analysis, thoughtful trade-offs, and architectural decisions that won’t be undone later.

Yet too few companies are building a transformation motion into how they operate. To be clear, we’re not advocating for a monolithic, command-and-control structure. That may work in some companies, but not in most. Instead, we see successful organizations installing a small transformation team to facilitate ongoing transparency and adaptability. This team supports business-owned solution teams that design changes against the multi-year outcomes agreed to by the leadership team. Solution teams test those changes in partnership with operations and scale them in a defined, repeatable model. The goal of the transformation team is to enable repeatability, coordination, and sustained value creation. 

In our experience, the emerging leaders have two speeds—run and change. Business functions play roles in both, with a focus on six critical areas:

• End-to-end process. Look across silos to reimagine how key sources of value deliver strategic and financial objectives.
• Solution team mobilization and drumbeat. Ensure solution teams are designed for testing and scaling, with clear steps for removing roadblocks and releasing funds.
• Data infrastructure and governance. Focus data efforts and investments on what drives the most value, not comprehensive fixes. Build capabilities to manage unstructured and synthetic data, and establish strong governance to ensure quality, reusability, and alignment with business priorities.
• Scaling. Commit to scaling change quickly and effectively across operations, geared to the unit of scaling (territories, factories, customers, etc.).
• Adoption. Build and sustain feedback loops, such as weekly adoption reporting, to support solution teams in scaling and visibility.
• Business and technology partnership health. Across the organization, increase visibility of enabling platforms, opportunities for reuse, and appropriate governance.

We believe this kind of transformation motion, focused on changing the business, must become a permanent defining characteristic of the modern enterprise. As AI disruption accelerates, companies will need to continuously balance “run the business” and “transform the business” objectives, often simultaneously. And the number of cross-functional, strategically important, and operationally complex issues continues to grow. How do you architect your processes for fully agentic workflows? How will you organize and govern the explosion of unstructured, AI-generated data? How will you lead and train an organization made up of humans, agents, and maybe even robots?

The questions will keep coming as AI evolves. Answering them won’t be a one-time effort. It will require a defined, ongoing transformation motion built into your operating model and a set of change capabilities built for the long haul.

From experimentation to enterprise transformation

AI isn’t just another tech wave. It’s a foundational shift in how work gets done and how value gets created (see Figure 1).

For most companies, the question is no longer how to use AI. It’s how to compete in a world where every competitor is using it too.

The winners won’t have the most pilots, the flashiest demos, or the biggest technology budgets. They’ll be making strategic choices, combined with the operational rigor to follow through.

One executive recently commented that after resetting unit cost goals for their entire core value chain and after one year working toward reimagining what work gets done and how it gets done in a more AI-native way, the company is now delivering double the EBIT margins of their competitors. 

Getting unstuck is a strategic imperative given the pace of technology development. The companies that act now will turn gen AI into real results. The rest will be catching up.