What is humanity’s greatest innovation of the industrial age—the steam engine? Penicillin? The internet?
These are all good candidates, yet without the productivity gains from another crucial innovation, the modern business company, it’s hard to imagine how these inventions could have improved the world at the same scale.
The trouble is that few industrial organizations have truly cracked the productivity code to maintain high performance over the long run. A recent analysis found that as of 2020, the average life span of a company on the S&P 500 index was down to 21 years, compared with 32 years in 1996.
A range of technologies, including most recently generative AI (gen AI), have promised to transform industrial production. But despite the real progress some companies have made in the Fourth Industrial Revolution, productivity growth remains stubbornly low across most of the world’s largest economies (Exhibit 1).
What’s missing are the organizational structures required to make technology truly work.
Large industrial organizations have traditionally focused on excellence in processes, often for the mass market, such as cars rolling off an assembly line. But today, standard work is much more automated. What’s left is increasingly nonstandard, yet most industrial organizations aren’t structured for nonstandard work—leaving them unable to change at the pace innovation requires.
The experiences of exceptional organizations show that capturing technology’s opportunities means making eight mutually reinforcing shifts that strengthen people, processes, and technology in tandem:
- from process excellence solely focused on repetitive tasks to also applying the principles of excellence to flexible, project-based work, turning the unusual into business as usual
- from control and compliance to empowerment, giving individuals the autonomy and information they need, often based on digital footprints, to work more effectively
- from gathering expertise in specialized silos to making expertise available to all via app stores, allowing workers to integrate highly accessible digital tools for tailored solutions
- from process masters to process reengineers so that rather than simply mastering current processes, managers continually rethink and redesign how tech-enabled processes can improve work
- from going to work to supervising control centers that are increasingly decoupled from physical production flows—opening new opportunities to restructure network strategies
- from Industry 4.0 to “Industry 4 U,” which yields better work environments via stronger technology support, improved ergonomic design, and an inspiring leadership culture
- from ad hoc learning to industrialized capability building to make rapid skill building with large numbers of employees a competitive advantage
- from building internal departments to orchestrating an ecosystem, redesigning how companies integrate knowledge and radical innovation into their value chains by rethinking collaboration models
Building on previous insights codified in methodologies such as agile and lean, the eight shifts collectively represent the greatest change since the matrix organization emerged in the wake of the mid-1960s space race. Meeting the challenge will stretch every critical enterprise capability our colleagues outlined in Rewired: The McKinsey Guide to Outcompeting in the Age of Digital and AI. The question leaders now face is how to prioritize the five actions they must take.
Turning the unusual into business as usual
Mass production of a standard product was the archetype for the modern enterprise. Standardization of products supported standardization of processes, culminating in management systems such as lean, which rely on elaborate codification of “standard work.”
Today, the context has changed. As automation and gen AI take over more tasks, human work increasingly centers on the sorts of “special projects” that once accounted for only a small share—perhaps 20 percent—of day-to-day work. The unusual is becoming business as usual. Yet governance still consists of heavy layers of (functional) processes on top of project management, yielding convoluted, slow, and resource-intensive decision making.
The challenge for business leaders is to flip the ratio: instead of deploying 80 percent of people in processes and 20 percent in projects, they can aim for 80 percent in projects. A few companies have successfully transitioned to a project-based organizational model, such as a major life sciences company that rebuilt its internal structure to support rapid growth. Crucially, the model balances stability and agility by ensuring that every employee has a home unit centered around their skills even as they are working within dynamic project teams whose composition is tailored to specific skill needs.
Within this structure, three additional factors proved critical:
- Build internal talent markets. Leaders can’t assemble effective teams unless they know who in the organization has the required skills. That means both building a skills profile for each employee and making it widely available within the organization. The company’s employee directory now provides detailed skills-based information, such as certification, training coursework, and recent project experience, which is automatically updated after a project ends.
- Set clear boundaries. Clear boundaries were established to operate within, consisting of company priorities, product strategy, project team mission, and short-term team goals. It was also important for senior management to provide rigorous oversight; in this way, they could identify opportunities for teams to share resources or join forces and thwart mission creep within teams.
- Develop achievement-oriented leadership. Company executives recognized that effective leadership of a project team required different skills from those traditionally encouraged in line leadership roles. Rather than directing the work of people with similar training but varying levels of experience, project leaders would need to coordinate individuals with deep expertise across many fields. The role of the leader therefore changed to emphasize what the project sought to achieve—the problems to be solved—along with why and when. The role of the team centered on how—collaborating to find the best solution.
Getting empowerment right
How do you plan in a world that is becoming more and more complex? Often, the counterintuitive answer is: you don’t. The humble baguette illustrates: every morning, despite countless potential obstacles, French bakeries sell millions of baguettes—all in what can appear to be a miracle of nonmanagement. Thousands of individual decisions organically converge to produce an essential good, with no need for a central baguette planning department.
However, too many large and complex organizations have yet to emulate the baguette example. They still seek to amass expertise in a traditional hierarchy, even though the assumption that the top of the organization generally knows better than the bottom was never more than partly true—particularly when it comes to understanding how operations really work. The trouble is that many organizations that tried employee empowerment failed and thus reverted to central control.
Getting empowerment right requires a balance: giving workers the independence they need to execute tasks effectively while being prescriptive about what those tasks are meant to achieve. “Independence” and “prescriptive” may sound contradictory, but as long as they are kept in close balance, each supports the other in generating more value (Exhibit 2). Together they call for a new level of transparency regarding both access to data and the data’s implications on results.
Guided by OKRs
Yet access to data is hardly the same as being informed. Frontline workers may now have tools to see exactly how productive their machinery is for different products, but that alone can’t determine how best to improve the line’s productivity—let alone its profitability. Instead, they need to understand how the organization’s strategy translates to the factory floor.
Leaders face the same question, but from the opposite perspective: the need to achieve strategic goals without the wasted effort involved in trying to dictate exactly how the goals are to be achieved.
Originally developed at Intel and expanded at Google, objectives and key results (OKRs) have emerged as an effective balancing mechanism, focusing the organization’s energy on outcomes—the “what”—while allowing greater flexibility on the “how.” The idea is to enable subunits within the organization to use the strategy to articulate goals and, crucially, the results that will matter in achieving them—thereby aligning what they do each day with the organization’s direction.
Defining—and achieving—OKRs calls for clear direction from senior leaders about the company’s strategic goals, as well as communication at all levels of the organization. Accordingly, an empowered organization requires more effort from senior management, not less. For the factory team looking for ways to improve the line’s profitability, OKRs would need to be tailored to the site, accounting for factors such as its position in the supply chain, its current capabilities, the training of its operators, and the needs of in-development products. Frontline workers and managers can then make intelligent trade-offs between maximizing profitability from current production and reducing potential costs for future changeovers.
Using activity-based analytics to drive business performance
The familiar sight of delivery trucks illustrates even more opportunities for empowerment. When a driver finishes a delivery, a quick scan tracks completion in the company’s IT system—which also knows who the driver is, where both the driver and vehicle are located, and which route the truck took to get there. The resulting digital footprint gives the company a very different way of managing tasks, controlling quality, and handling documentation.
It also helps employees, suggesting alternative routes when traffic is heavy or finding a different delivery location if the original is inaccessible. Should priorities suddenly change, the system can reprioritize the worker’s tasks automatically—so that the employee can more easily squeeze in a rush delivery and still finish the day on time.
This logistics example finds its equivalent in manufacturing execution systems (MESs) that are broadly installed in factory environments. Advanced versions can do even more, helping employees build their skills and, ultimately, advance on their career paths (Exhibit 3).
Take the example of a manufacturer of customized technology products. Company leaders learned that frontline assembly workers were often frustrated that when production problems arose, their work wasn’t tracked in enough detail for them to find (and fix) the underlying issues. They wanted a system where every line station would accurately log when the project arrived at the station and which assembly tasks were performed, and then see where workers needed more training to avoid rework. The leaders realized that building such a system would improve both product quality and employee retention, without requiring more time or attention from middle managers.
Indeed, technology can now assume many of the work allocation, planning, performance tracking, training, and issue resolution tasks that consume most of a middle manager’s day. The implication is that many middle manager roles could soon become very different, emphasizing process confirmation and team building as teams become increasingly self-managed (Exhibit 4).
Making expertise available to all via apps
In day-to-day life, knowledge that once took years to acquire is now available instantly, often through an app. But too many companies have been slow to follow this route. Early adopters show how using customized apps in manufacturing operations, for example, can help frontline employees see not only what they are doing but the value their work is creating. The challenge for companies is to create platforms that replicate these results wherever possible, via apps for everything from monitoring equipment and planning maintenance work orders to guiding defect resolution or extracting strategic insights for production planning.
That would mean investing in the standard prerequisites. Companies need data architects to build a flexible technology stack, data scientists to manage analytics, external partnerships to support the latest app development tools and licenses, and training programs to help the workforce use apps in solving everyday problems—optimizing inventory, debottlenecking production, or managing logistics.
It also means preparing for jobs to evolve, as traditional and tech-centric roles start to merge. Today, the chief information officer (CIO) is a crucial stakeholder for the COO, but tomorrow the COO could simply be the CIO. Likewise, other leaders would have a “major” in data architecture and a “minor” in their respective functional domains. In e-commerce, for example, category managers are increasingly data architects first, under the assumption that they can quickly learn about product categories as needed.
Finally, it means getting serious about building a culture of rapid, constant innovation. Software used to be something most companies would buy. Today, more companies build their own code, written by internal staff as part of a continuous improvement process—which, in the digital world, often boils down to the speed of app evolution. Building an app to track overall equipment effectiveness (OEE) for individual machines is not a one-time commitment. Instead, the app could even evolve toward self-improving as data gets richer and modeling becomes more sophisticated.
Remastering processes
A perennial problem for organizations is the mismatch between how a process is intended to function and how it actually functions (Exhibit 5). Continuous-improvement disciplines seek to reduce the gap, to create “process excellence.”
This problem disappears once a process is correctly designed and fully automated. The ultimate example is the so-called lights-out factory, one in which automation takes over production. Snack company Mondelēz has fully automated dough production in Beijing. Analytical models optimize quality, such as the texture of the dough, and a closed-loop system adjusts the machine settings in real time.
What does it take to go fully lights-out? Humans. Over the next decade, hundreds (if not thousands) of process reengineers will be needed. Chief process redesign officer may become the latest C-suite title, joining recent predecessors such as chief transformation officer or chief digital officer. And even after processes are reengineered, humans would likely still be essential to build and improve automation models and supervise the system from the control room.
That raises the question of where organizations could find the right talent, particularly at leadership levels.
Reshaping production networks with control centers
The Mondelēz example illustrates how automation is changing the factory’s role, but that’s only the start of the story. The COVID-19 pandemic opened everyone’s eyes to the feasibility of remote work. Once most manufacturing tasks—including ones that formerly had to be done in person, such as quality control—require access only to data rather than machinery, sites may start to lose a long-familiar feature: the control room.
If co-location with the production site is no longer necessary, where could control rooms be? Not just anywhere, or where labor is cheapest. The key criterion will likely be access to the right talent and skills, some of which are currently in short supply in developed economies. Technology capabilities are paramount. India, for example, is the world’s second-largest producer of graduates in science, technology, engineering, and math.1 It’s also a heavyweight in gen AI use. A recent global survey found that gen AI penetration was highest among organizations in India, at 81 percent, with Singapore ranking second at 63 percent and Spain third at 57 percent.
Companies’ network strategies thus face a refresh, with factors such as geopolitical resilience, language skills, and time zones all important to the mix. Existing shared-service centers can provide a starting point. Their experience in reengineering transactional processes—which increasingly are automated, leaving knowledge tasks organized into service hubs—gives them crucial expertise that reverses the previous outsourcing and offshoring model. Rather than standardize a process before handing it to the service center, it’s the service center that has the competence to standardize.
Industry 4.0 to Industry 4 U
By 2040, advanced manufacturing economies from China and South Korea to the European Union, United Kingdom, and United States are set to see contraction—or, at best, much reduced growth—in their active workforces as societies age. In the United States, for example, the ratio between people aged 20 to 24 and people over 55 in the manufacturing workforce dropped by 16 percent between 2014 and 2022.
Raising productivity growth will be essential to maintain standards of living in advanced economies while raising them in emerging ones. Investments in technologies such as advanced automation and gen AI, especially for routine and knowledge tasks that previously were done by humans, could unlock billions of dollars in economic activity.
Even more important, it would free up workers for higher-skilled work, where labor demand is already outstripping supply. Companies in especially affected sectors, such as US construction and manufacturing, are discovering the importance of an attractive work environment that provides opportunities to learn and grow. The technology that workers use becomes part of an employee-centric approach to digitization, one that seeks to fundamentally change the employee value proposition. “Industry 4.0” becomes “Industry 4 U.”
Industry 4 U is about giving workers the right tools, based on technologies that they’re using in their personal lives, along with a workplace that’s healthy and inspiring. Together, these elements add up to an engaging leadership culture. At a global oil and gas operator, this transition anchored a restructuring of the end-to-end employee experience so that the company could better attract and retain new workers—an especially difficult task in an era of increased environmental awareness and competition from the tech industry. On its own, the company couldn’t reverse climate change or technology shifts. But it could address long-standing on-the-job frustrations for a better working environment, especially during the critical first months of an employee’s career.
Listening to early-stage employees revealed a strong need for additional support, particularly in the renewable-energy business unit. The company responded with a comprehensive employee onboarding app that guided new colleagues through the first few weeks of work. Built in only three months, the app not only created a more positive and connected experience for new hires, reducing attrition, it also resulted in a new way of working for the technology team.
Industrializing capability building
For an experienced process engineer with a PhD in pharmacy, today’s definition of a successful career would likely mean leading the development of a handful of drugs that, after navigating a complex, high-risk regulatory environment, finally reach the market and are helping patients.
That definition is poised to change. What if instead of a handful of successfully launched drugs, that engineer could point to dozens—because deep learning models had eliminated much of the guesswork that led to development dead ends? And what if newer AI models had made process refinement faster and more accurate so that more products could be developed more cheaply?
McKinsey research estimated that by 2030, activities accounting for 30 percent of current work hours could be automated. Reaching this world would require the engineer to learn new skills, particularly in softer skills such as design thinking—critical for blending decades of hard-won expertise with the promise of AI. Leaders recognize this need: for integrating gen AI into their processes, more than half of employers said they plan to rely on internal capability building to meet their skills needs.
Traditionally, industrial organizations have relied on apprenticeship-based capability building. That model worked well enough for helping small numbers of workers develop well-understood skills, but it can’t meet the scale and speed needed for the rapidly evolving world of AI. The good news, however, is that the upskilling gap may now be shrinking because of easier access and lower cost of reskilling tools. Online courses and immersive-learning platforms make upskilling easily available on a self-service basis. Individual course segments can be slotted into day-to-day business, allowing for continuous learning. Gen AI even offers personalized coaching and advanced translation capabilities, making highly tailored trainings easy to access globally.
A global consumer goods company provides an illustration, having undertaken a two-year global digital transformation across its network of manufacturing plants and its supply chain.
Capability building formed the core of the transformation. Learning centers focused on innovation served as crucial support systems in building a company-wide academy that now leads everything from initial inspiration sessions to immersive capability building. The impact achieved was substantial: productivity increased by more than 15 percent across the network, for savings in the tens of millions of dollars. Additionally, the identification, sizing, and mapping of several hundred unique use cases paved the way for a comprehensive transformation road map. With more than 2,000 learners upskilled across the entire production and supply chain network, the company is now well situated for further improvement.
Orchestrating the ecosystem
In his seminal 1937 book, The Nature of the Firm, Ronald Coase argued that large corporations emerged because they reduced transaction costs: getting things done internally was significantly cheaper than outsourcing or buying services or goods externally. When data and expertise were expensive, having a planning department that understood the organization and could access the right information made economic sense. Today, the data can sit in the cloud, easily accessible to anyone—as can sophisticated planning applications that can import the data at a mouse click and produce a plan.
Companies now have a generational opportunity to reexamine the boundaries of their organization in answering the most fundamental question: What should we do, and what should we work with others to do? And as transaction costs keep falling—from faxes and the internet to social media and gen AI—where should the work take place? Understanding where the company creates the most value can reveal new opportunities not only to compete but also to collaborate more deeply, turning transactional relationships into genuine partnerships.
For a major utility network, the critical question was one of resilience, both operational and strategic: climate change is increasing stress on physical infrastructure even as the energy transition, new technologies, and geopolitical uncertainties are transforming how energy is created and distributed. Navigating these conflicting pressures led the company to realize that it would never be able to accumulate all the relevant expertise in-house: if it kept trying to do everything, it would do nothing well. At the same time, tapping into external services had never been more cost-effective.
This led the company to a major shift in mindsets. Rather than think of its suppliers as vendors to be given minimal information for maximum negotiating power, the company examined where data sharing would yield better-quality outcomes—and help the company focus its expertise. So, for example, careful sharing of pricing and cost data for types of distribution meant that suppliers could provide better economies of scale and delivery terms.
At the same time, entering a partnership means managing an inherent risk: overdependency. In many cases, this is linked to data access and ownership; conversely, keeping the right data can turn a potential vulnerability into a major strength. After all, Airbnb has data, but no beds; Google Maps has data, but no restaurants; and Uber has data, but no cars—yet each has disrupted an entire industry.
This data-centric model could help manufacturers tap into entirely new sources of growth, especially in industries where hardware may face diminishing returns. Rather than try to turn themselves into software providers, some manufacturers may generate more value under an app store model, with the physical product becoming more competitive because of the wider range of software it can support. This option requires resources—developing the store platform, identifying and protecting the most valuable data, and managing risk and compliance—but could be rewarding for manufacturers willing to commit.
Preparing for the eight shifts
The question for companies isn’t whether these shifts are coming. They’re already under way. What matters is to get started, focusing on five very concrete tasks:
- Understand and shape your role in the ecosystem. What do you do today better than anyone? That’s your current role in the ecosystem. But how will that look in five years? What are your blind spots, and how do you evolve to meet them head-on?
- Develop visionary and courageous leaders. What leadership profiles are currently thriving in your organization? What sets them apart from others? Are they the right ones to steer your company through the technological revolution? And how can you form a coalition with them?
- Define and invest in your tech stack. In five years, what technologies will you say you wish your organization had invested in today? Which ones will need integrating into your tech stack? Is your tech stack flexible—and are you ready to use that flexibility to accelerate innovation?
- Assess capability needs. Which skills of today are likely to remain relevant in the future? What new skills are coming, and how will your organization acquire them (upskill, hire, outsource)? How will capabilities affect your network footprint? Does your capability strategy incorporate the skills available in your ecosystem, not just on your payroll?
- Work in teams, not in pyramids. How are your teams set up today? How much of your work is dedicated to standard tasks to keep lines running, and how much is left for innovation and improvement? Where should the empowerment balance be between independence and prescriptiveness?
Individually, the actions companies face are not new. According to the Oxford English Dictionary, the title “chief executive officer” was first attested in 1914. Corporate training took hold in the late 19th century. Digital Equipment Corporation deployed a matrix organization in the early 1960s. The revolution is not in the individual steps: it’s in the questions that organizations must ask themselves as they consider how the eight shifts affect them. Successful leaders will shift from an evolutionary mindset to a revolutionary one and act now to fix their organizations so that technology can truly accelerate productivity growth.
