The new management playbook for AI: How to move faster and create more value

The companies that are truly innovating with AI aren’t winning just because of the tech they use—those tools are broadly available. Their advantage comes from how, and how fast, they apply technology to solving real business problems at scale. That requires new organizational capabilities that take time to build.

In the end, those capabilities become the true long-lasting competitive advantage that enables these companies to sustain a higher rate of innovation with technology.

We studied 20 companies that have consistently created significant economic value from their business transformation enabled by AI. These companies, many of which are profiled in the second edition of our seminal book, Rewired: How Leading Companies Win with Technology and AI, aren’t the wunderkind tech companies that are always in the headlines. They are large businesses that have invested the time to build up their tech and AI capabilities to successfully turn tech into value.

Game-changing impact with AI is already a reality

Our most recent State of AI report reveals that 94 percent of businesses have yet to create meaningful value.¹“The state of AI in 2026: On the road to ROI,” McKinsey, August 25, 2026. The math is unforgiving, and it underscores the skepticism that many boards and top teams have in making the level of investment necessary for successful business transformation with AI.

A small number of companies, however, have shown radically different outcomes. For each of the companies we studied, we reviewed the transformation road map and financial outcomes. We also interviewed select executives to gather additional details and insights. This analysis helped us answer important questions we are frequently asked about the economics of successful business transformation with tech and AI (exhibit).

These leading companies shared the following notable traits:

• They achieved game-changing bottom-line impact. We looked at the steady-state EBITDA achieved by these transformations after three years as a percentage of the EBITDA of the company. If the transformation was focused on a particular division or business unit, we focused our analysis at that level. The 20 companies in our sample improved their EBITDA by 20 percent on average.
• They focused on key economic leverage points. Two-thirds of the companies focused on three business domains or less. These business domains were always key economic leverage points where even small improvements by AI translated into material financial impact. For example, Freeport-McMoRan focused on process yield and throughput, a classic leverage point in mining. LATAM Airlines Group focused on passenger experience.
• They built AI systems, not point solutions. These companies didn’t approach their transformation by just pushing a single technology lever (like gen AI). They defined the problem to be solved and from there they designed an AI system to improve business performance. These systems were composed of different solutions and underlying technologies. Toyota used machine learning to optimize forecasting, agentic workflows to support supply planners, and digital workflows to provide supply chain transparency to dealers and customers, among other solutions. Just as important, these systems also contained nontechnology elements such as improving data quality, changing work practices, and realigning incentives.
• They became cash accretive quickly. These 20 companies became cash positive in one to two years on average. While they took longer to realize the full benefit (typically three to four years), they were smart in sequencing the build of their AI system to deliver low-hanging fruit that paid for the journey.
• They made substantial investments and achieved high returns. We measured the cash-on-cash return as annual incremental EBITDA divided by one-time cash investments. On average, for every $1 of investment, these companies realized $3 of incremental EBITDA. These extraordinary returns are the result of focusing efforts on economic leverage points rather than pursuing a “thousand flowers bloom” approach to AI. These companies made substantial investments in transforming their business with AI, often in the range of $50 million to more than $200 million.

Interestingly, despite notching such amazing successes, these companies are still at it. Improving business performance with technology never stops. They typically go through waves of transformation that last two to four years. Freeport had its first large-scale AI breakthrough in 2018 when it created step-change productivity improvements in its copper concentrators. Three years later, leveraging that newly developed “muscle,” Freeport repeated its prowess in leaching.²Leaching is the process of extracting copper from ore through the judicious use of chemicals.

Building the system: The capabilities needed to transform businesses with tech and AI

The ability of an organization to harness technology for business value is the competitive advantage, not the technology itself. We identified six core capabilities that successful companies build.

These six capabilities do not operate independently; they reinforce each other. They also take time to build. Our four exemplar companies have been at this for more than five years, and they are still at it. No company becomes an overnight AI success. But once a company has developed these capabilities, it has a real competitive moat.

A C-suite team that gets AI

A C-suite that understands AI is the most significant driver of success.

Successful top management teams understand that to create value with technology, they must focus their AI efforts where it matters, and they must use AI in a competitively differentiated way. Doing so requires strategic creativity to see what others do not. That’s hard.

At DBS Bank, the top team spent significant time in Silicon Valley and with digital natives to understand what truly differentiated high-performing tech firms—not tools, but capabilities: modern engineering practices, platform-based architectures, data at scale, agile ways of working, and a culture of continuous experimentation. This wasn’t a superficial benchmarking exercise; it led to a fundamental reframing of DBS’s technology operating model. The result was a multiyear commitment to build these capabilities systematically so that technology could consistently drive speed, innovation, and customer value.

This work cannot be assigned to a single executive or delegated further down the organization. This is the hard work that the top team must undertake, together.

Leaders on the front lines of change who have tech and AI muscles

Because AI is so deeply embedded in a company’s operations and workflows, harvesting value from AI requires leaders who combine deep domain expertise, an understanding of technology and data, and an ability to orchestrate end-to-end change.

When Freeport decided to go all-in on its leaching optimization system, it tapped a leader with both experience in operationalizing AI-based solutions and credibility in processing operations. It relocated him to be the general manager of its leaching operation, making him effectively the “domain owner” for this critical part of the business. As effectively the domain owner, he led the development of the AI system, integrated the different technology solutions into business operations, and was ultimately accountable for the business outcomes.

This is a consistent finding in our success stories: There is always a senior business leader³Sometimes the model can be two-in-a-box with a business and technology leader. who leads and integrates business, technology, and the change management aspects to achieve breakthrough outcomes. Of all the AI upskilling programs a company might launch, none will be more strategic and impactful than developing tech-capable business leaders.⁴Dana Maor, Eric Lamarre, and Kate Smaje, “Building the AI muscle of your business leaders,” McKinsey, December 1, 2025. This is where companies that are leading with AI today are devoting serious effort.

Rewired, Second Edition

This updated edition offers brand-new insights into cutting-edge AI solutions—and what it takes to implement them—as well as the new economics of digital and AI transformations.

Learn more

An operating model built for speed

Rewired organizations embed tech delivery capabilities in the business to drive more effective and faster innovation cycles. They build platform capabilities to maximize reuse and support faster business innovation. DBS Bank organizes work around end-to-end customer journeys such as opening an account, buying a home, or securing small business financing. Cross-functional teams are aligned to these journeys and are accountable for delivering seamless customer outcomes across channels and products. Rather than working within functional silos, these teams orchestrate capabilities from the underlying platforms (more on platforms below) to solve specific customer needs. The benefit is sharper customer focus with faster cycle time, better conversion, and increased customer satisfaction.

The bank is also structured around a set of enterprise platforms such as payments, customer data, onboarding, and credit. Each platform brings together business, technology, and operations and is funded as a long-term asset rather than a series of projects. These platforms build reusable capabilities—APIs, data assets, and services—that can be leveraged across multiple parts of the bank. The benefit is scale and speed.

This is what we call a distributed operating model because teams close to the business can build, test, and improve solutions iteratively. Every success story we have documented has a version of this operating model, yet only 10 percent of companies have adopted one.⁵“A new operating model for a new world,” McKinsey, June 18, 2025. Why? Rearchitecting a company around such a model requires vision and resolve.

‘Technology as a platform’ to power the enterprise

When LATAM Airlines was embarking on its transformation journey, it made a deliberate decision to build customer-facing capabilities separately from its legacy core and in the cloud. This was not simply a technology choice; it was a strategic architectural commitment. The new digital organization was designed to be API-first and modular from the outset, with reusable components and open integration layers.

Because solutions were built API-first and modularly, they could be reused across channels and geographies. An order-change capability developed for the web could be deployed to the app, contact center, or WhatsApp without rebuilding core logic. When scaling to new markets with different payment systems and regulatory constraints, LATAM used abstraction layers to integrate or swap external systems without destabilizing the platform. What could have become a patchwork of local customizations remained coherent.

Well-architected technology platforms accelerate everything: product launches, automation, data reuse, AI deployment, resilience, and more. Weak platforms slow everything down and tax the business with hidden costs through complexity, fragility, long cycle times, and dependence on a few heroic individuals who “know how the system works.”

The shift from siloed systems to platforms is one of the most important strategic capabilities that rewired companies have built. If you aspire to be a CEO—or remain an effective one—you must treat your technology platforms with the same rigor and ownership as your business strategy or succession plan.

Data that’s easy to consume

In 2018, it took DBS 15 to 18 months to develop and deploy AI models. Data was stored in different silos across the business, which meant it took months just to locate the data and get access permissions. Even when teams did get access, discovering how the data was structured and assessing its quality became a project of its own. Eventually teams managed to build one-off data pipelines that couldn’t easily be reused by other teams.

That just wouldn’t do. So, DBS embarked on a journey to develop a unified data platform as well as an AI platform, automate data access controls, and make data easy to both discover with metadata and to consume through data products. It aligned its data governance around the leadership of its different business domains, ensuring data was really managed as an enterprise asset.

Building this capability paid off. By 2023, it took just two to three months to deploy AI models. This ease of data consumption became central to unlocking an estimated more than 1 billion Singapore dollars (nearly US $772 million) in value generated from AI.

Designed for scale from the beginning

AI systems create value only when they are adopted and scaled. That may sound obvious, yet it remains one of the hardest challenges.

Adoption often fails because adjacent upstream and downstream processes are left unchanged. When Freeport developed an AI pilot to boost copper recovery through leaching, it created data and sensor standards so that the data could be easily ingested into a cloud data platform built for an earlier program and then reliably modeled in a consistent way across all of its assets.

Scaling is a different, but equally difficult, challenge. Expanding AI solutions quickly and economically across markets, factories, customer segments, or product lines requires modular solution architectures and a well-choreographed “dance” between central teams and the receiving units. Freeport solved this problem by recognizing that 60 percent of the AI system could be reused across plants while 40 percent would require local adaptation. This informed how to organize the central team that maintained the shared assets and the field team that ensured local adaptation.

The capability of an organization to manage these adoption and scaling challenges gets built over time as the company learns. Eventually, a success playbook emerges inside the company and continues to evolve with new technologies such as agentic AI.

Foundational capabilities are the secret to building greater speed

The companies that pull ahead do not treat early AI wins as end points. They know that competitors will eventually replicate their success. These companies constantly improve their AI systems with better data, more sophisticated models, real-time responsiveness, deeper orchestration across related workflows, richer customer/user experiences, and so on.

The companies that win that innovation race have capabilities that grow as technology and managerial practices advance (table).

The capabilities of each stage build on those developed in the previous one. In other words, don’t imagine that your company can skip to stage 3 without having mastered stage 2 capabilities. That’s why the companies that have historically led their industries in the strategic use of technology tend to be those leading again in the age of AI. They have the muscles to harness the value of AI!

The companies profiled in this article have largely mastered stage 2 and are now actively developing stage 3 capabilities. LATAM Airlines is very advanced in the adoption of agentic software development, increasing its effective development capacity by 50 percent while reducing team size by 30 to 50 percent.

We are in the early innings of stage 3. The next few years will see fast maturing of these new agentic capabilities. The companies that master them will compound even more advantage over peers that don’t.

The question is no longer whether AI will reshape your industry—it already is. The real question is whether your organization is building the capabilities to shape that future or simply reacting to it. The companies pulling ahead are investing now to build the muscles that allow them to repeatedly turn technology into business value.