AI is everywhere, with revolutionary potential for the world of work. Yet the transformation necessary to realize this potential is lagging. Most organizations have adopted AI tools, but few have realized meaningful changes in how people work to unlock the most desirable value pools: the areas of the business where AI delivers the greatest and fastest returns.
A recent BCG study reveals that although most companies are experimenting with AI, only about 5% are generating value at scale—and nearly 60% report little or no impact to date. Many see this gap and double-down on adopting more AI tools, rather than building the human capabilities that translate adoption into business value.
Traditional training is not enough to meet the scale and urgency of the GenAI moment. Winning companies are building integrated, enterprise-wide enablement systems that embed AI into how people think, work, and lead. And their efforts are focused on the parts of the business where the transformative potential of AI-enabled work is greatest.
Subscribe to our Artificial Intelligence E-Alert.
Realizing AI’s full value potential requires a step change in an organization’s capabilities—the mix of technical fluency and human skills supported by behaviors, mindsets, and systems that help people work differently, and more effectively, with AI.
BCG U helps organizations build the capabilities their people need to thrive in the age of AI. We’ve coached executives through complex AI-related mindset shifts, helped global organizations build AI fluency at scale, and designed hands-on labs and behavioral nudges delivered in the flow of work that accelerate how people use generative AI (GenAI) to transform how work gets done, translating into business value.
Across these experiences, we’ve seen that capability-building only drives impact when organizations move beyond traditional training programs. What works best is a three-part learning progression that produces lasting behavior change:
- Foundational. Impart new knowledge through key concepts, frameworks, vocabulary, and “a-ha” moments.
- Applied. Turn knowing into doing through on-the-job practice, tied directly to real workflows.
- Embedded. Turn doing into habit by codifying new practices into ways of working, role expectations, support structures, and incentives.
Traditional training fails to deliver impact because it stops at the first step. Yet moving through all three stages is complex and resource-intensive. That’s exactly why organizations must prioritize capability-building in the areas of the business where AI’s full value potential, once realized, can generate the highest return on investment.
In this article we offer insights about building AI capabilities in the workplace and lessons we have learned helping clients navigate the journey from AI experimentation to impact.
Ground AI Enablement in a Clear Context
To drive impact when building AI capabilities, the context for the effort must be clearly defined by identifying high-potential value pools and redesigning the related workflows. The organization also must understand its existing AI capabilities so it can build from a realistic starting point.
Most organizations—and even teams and workflows within them—sit at different stages of AI maturity, often straddling more than one stage at a time. (See Exhibit 1.) In practice, building from the starting point means understanding where you are on your AI journey. The next step is to define where you want to be—at the team or workflow level. Only then can you build the specific capabilities required to move up the curve and unlock value pools with the highest potential.
We worked with a leading European retail bank as it set out to reimagine its lending operations, embedding GenAI into end-to-end workflows to improve both efficiency and customer experience. The organization radically simplified its core lending processes and introduced an “Ops AI Agent” to automate document validation, plausibility checks, and data transfers.
The results were transformative: the bank achieved more than a 50% productivity gain across lending workflows, cut manual processing time by 70%, and reduced approval cycles from days to under 30 minutes in some cases. Employees were able to focus on higher-value work while customers benefited from faster, more transparent interactions.
Several key activities drove this transformation. The bank began by focusing on a clear value pool with the highest potential business impact—in this case, lending operations, where automation could reduce costs and improve customer experience, They then redesigned workflows around that opportunity, rebuilding lending processes to enable AI-driven efficiency, identifying friction points, redesigning handoffs, and codifying new ways of working. Finally, the bank identified its starting point by assessing its AI capabilities, defined a bold future-state vision, and developed a roadmap to build the capabilities necessary to capture this value.
Develop Role-Specific Capabilities
AI transformation is an all-hands-on-deck endeavor. But not all hands have the same type of involvement. To optimize learning, our design philosophy at BCG U follows a roles-based approach targeting four AI-related archetypes:
- Shapers: Executives and senior leaders who set the AI vision for the organization
- Leaders: Managers who create the conditions for AI to scale
- Transformers: Team leads who rewire workflows to incorporate new AI tools
- Frontline Contributors: Individuals who use AI tools in their daily work
Our experience has shown that if you build fluency and confidence tailored to each role’s unique demands, momentum follows. (See Exhibit 2.)
For example, a global company in fast-moving consumer goods (FMCG) recognized early that GenAI would reshape competition across marketing, supply chain, R&D, and product innovation. Despite some early experimentation with various tools, progress was uneven and impact remained fragmented. Leaders made a bold bet: that rapid AI capability-building across the company would lead to competitive advantage.
Together, we designed a custom AI Accelerator with learning journeys for key personas that reflect the archetypes:
- Strategic persona. Senior leaders who need to motivate, inspire, and model AI use. Their journey focuses on building a shared narrative and practicing how to lead teams through change. Leaders engage in live virtual sessions, case study discussions, and immersive labs where they apply GenAI to real business challenges.
- Implementation persona. Managers and frontline employees who need to embed AI into their daily work. Their journey emphasizes hands-on simulations and use cases tied directly to workflows in marketing, supply chain, and R&D. This enables individuals to move from learning AI foundations to embedding new habits.
To ensure the program feels authentic, we built custom lessons that spotlight transformation stories from within the company to showcase early wins, surface lessons learned, and highlight peers already reimagining how AI drives value.
The AI Accelerator has been deployed to several thousand employees (and counting). As a result, the company is seeing not just an increase in the use of AI tools but also measurable gains in productivity and confidence. Employees feel a clear sense of momentum: through the program, they experience a shared language around AI and its benefits, understand how they can incorporate AI tools into their work, and grasp the business rationale behind the rollout. We have observed consistent behavior change, which is leading business units across the global organization to become more engaged in the AI transformation overall.
As another example, we recently partnered with a global biopharmaceuticals company to put this approach into practice, segmenting over 100,000 employees based on the four AI archetypes, each with a tailored learning journey. Leaders received vision-setting support and role modeling playbooks. People closest to the frontlines engaged in hands-on simulations and applied use case labs. Ultimately, the business increased adoption of AI tools from roughly 20% to nearly 90%.
These experiences show that thoughtfully tailored learning journeys can help turn curiosity into participation, participation into sustained behavior change, and behavior change into concrete business benefits.
Leaders Should Share—and Model—a Unified Vision
The most effective AI transformations begin when leaders are equipped to lead from the front, modeling the behaviors they want to see across the organization. Yet sometimes different groups in an organization speak different languages. For modeling to be effective, leaders must be on the same page about the transformation—from the process to the goals. Modeling also requires more than foundational understanding; it requires making desired behaviors explicit, practicing those behaviors visibly, and reinforcing them through clear accountability structures.
In the words of Kaye Foster, a BCG executive coach and former CHRO at Johnson & Johnson, “We assume that if leaders know what to do, they’ll do it. That’s just not true. Most of the time, leaders revert to the behaviors that are reinforced, not the ones they learn in training.”
We worked with one client’s executive team to build their AI fluency, enabling them to steer their organization’s transformation in a coherent, compelling way. Executives found it difficult to talk about technical issues, while technology leaders struggled to frame solutions in terms of business value. As a result, discussions between these groups of leaders often focused on point solutions instead of the foundational investments needed to transform the business.
We brought the leaders together for an immersive learning experience championed by the CEO. The summit blended the latest in AI technology education with business strategy, leadership, and execution. It provided space for the executive team to step back, challenge old assumptions, and engage in honest cross-functional dialogue. Leaders worked through actual use cases and confronted the realities of transforming a large, legacy organization under intense competitive pressure. Discussions centered around business transformation and impact.
The impact was immediate: both business and technology leaders had more confidence and a shared language to guide decision making. Executives had more technical fluency, and tech leaders had deeper perspective on business goals. Together, they rallied around the core belief that there is no AI strategy, only business strategy powered by AI. This shared understanding has evolved fragmented conversations into strategic, cross-functional problem-solving focused on scaling value. By setting the tone from the top, these leaders modeled the behaviors they expected across the organization: curiosity, collaboration, and continuous learning. This visible commitment to change transformed alignment into action.
Trust and Motivation Are Key to AI Transformation
AI rollouts can elicit emotional responses. Some people are deeply skeptical of AI and particularly wary of how it may impact their job security. Others have concerns about the quality of AI outputs or how to use new tools responsibly. For AI transformation programs to succeed, organizations need to engage with those concerns head-on.
According to Brad Strock, a BCG executive coach and former CIO of PayPal, “The real challenge with AI isn’t the technology—it’s getting people to trust it. If you don’t build trust first, no AI initiative will succeed.”
Recently, we partnered with BCG X to help a global technology company design and pilot an enablement program aimed at accelerating GenAI adoption among its software engineering teams. Developers initially approached the pilot with skepticism and uncertainty about how AI would affect their roles, and their reticence manifested as low engagement with any AI-related initiatives.
To build trust and reduce trepidation, we first upskilled managers in GenAI fluency and change management skills, equipping them to set expectations, coach their teams, model responsible AI usage, and address concerns with empathy while reinforcing a clear business case for change. We also worked with managers to guide their teams through adoption sprints that combined hands-on experimentation with peer reviews of AI prompts and small-group labs.
This combination of top-down leadership support and grassroots team engagement created a safe environment for employees to try new workflows, share successes, and learn from setbacks. It also demonstrated that their feelings, questions, and concerns were understood by management. As a result, teams participating in the pilot not only adopted AI tools at higher rates but also reported stronger confidence and motivation—seeing a potentially threatening technology as a source of empowerment and pride.
Ultimately, emotions are always present in moments of transformation. The most successful programs don’t avoid them—they acknowledge them: anticipating skepticism, creating room for dialogue, and empowering individuals to participate with purpose. Only then can capability translate into confidence, and confidence into action.
Measure What Really Matters: Value
Measuring an organization’s level of AI adoption is essential, but it’s only the starting point. True success with AI comes from measuring how new capabilities unlock value pools across the business.
In the words of Marty Gilbert, BCG executive coach and former senior executive at Kaiser Permanente, “One of the biggest mistakes in digital transformation is thinking that adoption is the goal. It is not. The goal is making the technology valuable for the people who use it.”
The people closest to the frontlines may adopt an AI tool without maximizing its benefits, or they may leave value on the table by adopting a tool only temporarily. To encourage success, leaders should focus on metrics that show business impact and tangible results—like faster time to market, improved decision making, cost savings, or better customer outcomes.
At BCG U, we approach measurement as part of program design. We start by defining the desired business outcomes, then we delve into the behaviors that need shifting, the skills that need building, and the learning experiences required to enable them. (See Exhibit 3.) Once the program is in-market, we then trace impact in reverse order: capturing the learner experience, testing for knowledge gained, observing behavior change, and ultimately tracking whether the program delivered the intended business outcomes.
Traditional learning metrics like program completion rates or quiz scores fall short of capturing real transformation value. Instead, businesses should track depth and variety of tool usage, experimentation rates, observed behavior change (for example, better prompting or improved quality assurance), and tangible business outcomes including time saved, errors reduced, and decisions accelerated.
We recently partnered with a leading financial services group to upskill hundreds of employees in GenAI. In addition to measuring adoption, we analyzed the volume of new use cases generated, the rate of experimentation, and the business value from time saved and errors reduced. The results were impressive: 98% of learners generated new use case ideas, 80% applied learnings directly to managing projects, and 85% reported using AI more frequently at work.
These metrics demonstrated a deeper capability shift and the beginning of lasting business impact, not just surface-level adoption. Learners were using AI daily and coaching others, creating momentum, and delivering results across the organization. Had we only measured adoption rates of AI tools, we would have missed the more significant value the program was unlocking.
Developing AI Capabilities to Meet the Moment
The GenAI moment requires transformation on many levels. Organizations can reap the greatest rewards by deploying an integrated, enterprise-wide AI enablement system that empowers their employees and is tightly linked to business value. This should include personalized, role-based learning journeys that evolve with an organization’s AI maturity; behavioral interventions that help employees reshape how work gets done; technology and change intelligence tools that monitor usage, diagnose friction points, and enable real-time course correction; and end-to-end measurement systems that track capability shifts, behavior change, and ROI.
Taken together, these elements form the foundation of a next-generation AI enablement engine, one that moves organizations beyond AI adoption to build and scale the capabilities required to unlock significant and lasting business impact.
