Today, a BCG team redesigning a client’s organization can collaborate with an AI agent that encodes 63 years of BCG’s expertise, accelerating the initial analysis and freeing experts to spend more time with the client on decisions, tradeoffs, and implementation of the redesign.
That agent is one of several hundred client-ready AI assets built by BCG’s technologists and the experts who know the subject best. In parallel, BCG is undertaking the largest upskilling effort in its history: nearly all of the firm’s 33,000-plus employees have completed the first level of its AI Fluency Ladder, and more than 20% have reached the highly technical fourth level. Beyond upskilling, the larger ambition is to combine AI with BCG’s expertise to change how its work gets done: where analysis can be accelerated, where human judgment matters most, and how teams can focus on activities that create the most value. The result is a human-led organization strongly augmented by AI—ready to write the next chapter.
None of this was in place in 2022.
BCG had been strengthening its AI capabilities for a decade. But when generative AI entered the mainstream at the end of 2022, the firm faced the same questions that many of its clients did. No one knew how quickly the technology would develop, which models would prevail, or where the most valuable applications would emerge. Even so, BCG made an early decision to learn by doing rather than wait for the technology to mature, building on the capabilities it had already developed.
“Nobody knew where we would be in four years, and no one knows exactly where we’ll be in another four,” said BCG’s CEO Christoph Schweizer. “What we knew early on was that we wouldn’t wait and see.”
From the beginning, BCG invested in its people as well as in the technology, redesigned workflows rather than simply automating tasks, and stayed flexible enough to change course as needed. That commitment continues to grow. BCG plans to invest more than $3 billion in AI over the next three years.
Much of the journey has followed a framework that BCG calls Deploy, Reshape, and Invent. Deploy puts AI into widespread use while building the fluency and safeguards necessary to use it well. Reshape redesigns workflows, knowledge, and the organization around what AI makes possible. Invent explores how AI can expand the outcomes that BCG delivers and open new areas of client value.
The three have developed in parallel rather than in sequence, informing one another. And the story of each continues to evolve. BCG’s experience as “client zero” shows that the familiar principles of transformation still matter, even when the underlying technology is evolving fast. It also points to a larger lesson: adopting AI is just one part of the journey. Creating sustained competitive advantage requires making choices about where AI can genuinely change the work, combining it with proprietary and differentiated expertise and context, and redesigning the organization around it.
Deploy: From Access to New Capabilities
From the outset, BCG made AI broadly accessible. Adoption mattered because it was clear that many of the most useful applications would be discovered by people in case teams and different functions using AI in their day-to-day work, rather than by a central team. But adoption was only a starting point.
Through Amplify Impact, its firmwide AI enablement program, BCG combined access to tools with structured learning, peer support, role-specific enablement, and responsible-use standards—creating room to experiment while raising the quality of AI use.
Experimentation spread quickly. By early 2026, more than 90% of BCGers were using AI, and roughly 60% were using it habitually. More than 11,000 people had created approximately 45,000 custom GPTs. That did not translate into 45,000 durable applications, but it showed how broadly experimentation had spread. In BCG’s 2026 People Survey, 81% of respondents said that generative AI was helping them create more value from their work, and 73% said that it made their work more enjoyable.
Peer learning became part of the model. “When a colleague sees someone they respect do something remarkable with AI, it’s more persuasive than any training program you can run,” said Alicia Pittman, BCG’s global people chair.
As the technology advanced, the firm added more structured ways to build proficiency. For example, its AI Fluency Ladder is tool-agnostic and grounded in real BCG work. Its premise is simple: having access to a tool and being able to use it well are different things, and different roles require different levels of fluency.
The Fluency Ladder therefore measures technical proficiency. It helps develop people’s ability to know when AI is useful, how to challenge its output, and how to combine it with expertise and judgment. To leaders’ surprise, nearly 100% of employees voluntarily achieved level 1 after four months, and more than 20% completed the highly technical level 4 training and certification. Word spread beyond BCG, leading BCG U to turn the Fluency Ladder into a product for clients that quickly became one of its most in-demand programs.
Responsible use of AI has developed in tandem with its increased capabilities. In 2022, BCG published one of the industry’s first AI codes of conduct. People remain accountable for AI-generated output, and client applications require appropriate transparency and safeguards. BCG further codified these expectations for all BCGers into BCG’s AI Core Standard commitments.
The firm’s ambition has moved beyond individual productivity. BCG has been building AI-native case teams, rewiring how teams function rather than simply layering AI onto an existing workplan.
An AI-native case team works with the same tools and from a shared case knowledge base that collects documents, meeting notes, and transcripts. Context no longer needs to pass manually from person to person. Consultants use AI for a first pass and spend more time testing and deepening the analysis. Project leaders increasingly orchestrate people and AI as an integrated team, freeing up more time to work alongside clients on decisions, upskilling, and the change required to turn recommendations into impact.
New team rhythms support that shift: AI norms are set at kickoff, teams conduct regular retrospectives, and time is reserved for people to learn together. Three months into the program, nearly 1,000 managers had completed AI-first team bootcamps.
Early experience has also revealed the limits of a simple productivity story. AI-native teams can reach first answers faster and explore more alternatives. But more output can create more review work for senior team members, and faster analysis does not mean that a client can make decisions or absorb organizational change at the same pace.
That tension goes to the heart of Reshape. The opportunity is not simply to make existing work faster, but also to rethink where AI changes the workflow, where expertise and judgment matter most, and how the surrounding organization needs to adapt.
Reshape: Redesigning the Work—and the System Around It
At BCG, seizing that opportunity has meant asking a more fundamental question in connection with the push for adoption: where should AI change the work itself?
The objective is to identify where AI can create differentiated value, where BCG’s expertise and context can strengthen the answer, and where human judgment should remain decisive: framing ambiguous problems, making tradeoffs, and helping clients turn analysis into lasting change.
That objective raised an organizational challenge. BCG is highly decentralized. Client teams and practices—areas of expertise organized by industry as well as by capabilities that cut across industries—are closest to the problems that clients are trying to solve. Their proximity enables ideas to emerge quickly and is one of the firm’s core strengths. But it can also create fragmentation. Similar solutions may be built independently, and something designed for one engagement may not be suitable for reuse across others.
“The tension is how to harvest the bottom-up innovation and the energy that comes with it, while still building the guardrails and scale you need,” said Teresa Espírito Santo, BCG’s global chief of staff.
BCG therefore built a common foundation to help useful innovation scale across the firm. Teams can build, discover, reuse, and improve AI capabilities developed anywhere within BCG without rebuilding the underlying technology.
Central to that effort is a program that turns BCG’s methodologies and proprietary data into reusable, AI-enabled workflows for client work. Ideas still originate close to the problem—in practices, regions, and case teams—but common architecture and standards make it easier to take the best of those ideas further.
This is increasingly happening at scale. BCGers used the assets built in the first wave of the program in more than 1,500 cases and proposals within the first few months.
It is critical for that foundation to stay flexible. When BCG first recognized a need for technology to connect AI models and internal data, no external solutions that fit the need, so the firm invested significantly in building its own. As external tools caught up, BCG made difficult choices to abandon bespoke programs in favor of superior external technology. In a market moving this quickly, the ability to adapt matters as much as getting the initial choice right.
As foundation models become more widely available, BCG does not expect the model itself to be the lasting source of advantage. Increasingly, advantage will shift to what surrounds it: proprietary knowledge, client context, and accumulated experience.
BCG has built a context layer that brings that knowledge into the flow of work through persistent case context and practice-level knowledge graphs. The result is AI that can be grounded in a client and case, integrating BCG’s accumulated knowledge and complementing its human expertise. As agents take on more work, evaluations will monitor their performance, and human accountability will remain in place.
This emerging approach creates a second-order challenge: as agents do more, organizations have to decide who owns them, who keeps them current, and who is accountable for their output.
These questions are already reshaping BCG’s knowledge and research organization. Increasingly, domain experts own AI-enabled workflows, while engineering teams build and run the technology. This is BCG’s 10-20-70 principle applied internally: 10% of efforts should focus on algorithms, and 20% on technology and data. But the bulk of the transformation—70%—is about people, processes, and ways of working.
For clients, exposure to this principle may be one of the most important benefits they gain from BCG’s experience. The technology matters, but the harder questions are often organizational: which workflows are worth reshaping, who owns them, and where human judgment remains essential.
Teams can find reusable solutions in BCG’s Applied AI Marketplace, a single internal destination for trusted agents and other digital assets. They also work more and more frequently with agents in everyday collaboration channels such as email and Slack, where agents take on longer-running work, surface questions, flag assumptions, and ask for guidance as they work.
One example is BCG’s Operating Model Agent, which combines AI with BCG’s methodologies and proprietary data to support aspects of organization design and transformation. Internal assessments have found cycle-time improvements of two to three times for some of the work, with expert review and judgment retained. The goal is to move expert time toward higher-value decisions and implementation.
Another example is a cost diagnostic agent. A team submits a company name and a question, and within an hour it receives peer benchmarks, the client’s P&L mapped onto BCG’s cost taxonomy, savings sized by category, and a list of the BCG experts best suited to pull in.
The initial diagnostic is only an entry point. The greater value comes as the work continues: teams can test hypotheses more quickly, rerun analyses as new information arrives, and keep context alive across the engagement. As a result, the team can devote more of its time to consequential work with clients—making choices, resolving tradeoffs, and turning the analysis into change. One goal of this approach is to arrive better prepared on day one, but an even more important objective is to sustain a faster, richer problem-solving loop through implementation.
Invent: Expanding What BCG Can Deliver
Deploy changes how people use AI. Reshape changes how work gets done. Invent explores what new forms of value AI can make possible.
For BCG, that means exploring where AI can expand the outcomes that the firm helps clients achieve—from new products and solutions to new ways of delivering transformation and staying involved in execution.
BCG has spent years building capabilities in technology, data, and design, alongside its traditional strengths in strategy and sector expertise. AI enables the firm to bring those capabilities together in new ways.
One example is Deep Customer Engagement AI by BCG X. Developed by BCG X, the firm’s AI build and design division, this product combines agentic AI, data, technology, and operating-model change to transform how companies manage customer relationships. The result is a technology-enabled solution embedded in how the business operates.
AI is also changing how BCG addresses large-scale transformation. The firm is testing AI-first offerings that combine expertise with new approaches to delivery, allowing teams to stay closer to implementation while changing the economics and speed of the work.
In ERP or other large system migrations, for example, AI can help teams make complex work easier to scale and find new ways to integrate solutions. “What if I can do this not with 100 people but with 20, and not in 10 years but in 5?” said Nicolas De Bellefonds, a BCG managing director and senior partner.
As a result, AI can change the economics of delivery and the deployment of human expertise—enabling transformations that were previously too cumbersome or uneconomic to become feasible and allowing expert teams to stay involved further into implementation.
This is where BCG’s combination of capabilities matters. The firm’s role is to combine strategic judgment, deep expertise, technology build, and transformation execution through new delivery capabilities. And it does this while enlisting the support of top AI frontier labs, hyperscalers, and other tech players as needed.
AI is expanding both what BCG delivers and how it delivers it, turning expertise into reusable capabilities, reshaping end-to-end workflows, and allowing teams to remain closer to the work that ultimately determines whether a transformation succeeds.
A Living Transformation
BCG’s experience as client zero does not translate into a finished AI playbook. The technology is moving too quickly for that. However, from BCG’s own transformation—and the hundreds of others that it has worked on with clients—several principles have become clear:
- Move early, but be willing to change course. Some early technology choices will inevitably evolve or need to be replaced as the market develops.
- Broad access matters only if capability rises with it. Adoption alone does not create sustained value. People need the fluency to use AI well and to challenge its output.
- Let innovation emerge close to the problem, but create the infrastructure to scale what works. Decentralized experimentation generates ideas; common architecture, standards, and governance allow the best of them to travel.
- Treat proprietary knowledge and context as strategic assets. As foundation models become more accessible, differentiation will increasingly come from what surrounds them.
- Redesign workflows and roles, not just individual tasks. The largest opportunity arises when AI changes how work is organized and where people spend their time.
- Keep human accountability explicit. As agents take on more work, human ownership, review, and judgment become more important, not less.
BCG’s own research illustrates why the transformation matters. Its 2026 Applied AI Index found that only 7.5% of companies are future-built, meaning that they generate sustained value and competitive advantage from AI. A separate study of CEOs found that high performers were roughly seven times as likely as others to redesign workflows and reshape the business end-to-end with AI.
BCG’s internal experience adds an organizational dimension to those findings. The hardest choices are about people, workflows, and the business: how fast to move, where human judgment remains critical, and what to standardize versus leave decentralized.
Leaders need to make those choices deliberately—and to recognize that the operating model, talent model, and economics of the business must evolve alongside the technology. To achieve transformative impact, companies will need to blend strategic clarity with applied AI across their business.
Some of these transformation decisions are familiar, but AI is forcing organizations to confront them at a new speed and scale. As Christoph Schweizer put it: “It will be as much a change management effort as a tech one. It has to be owned at the top.”
