When I met with leaders from one of the largest public utilities companies in the world for an artificial intelligence strategy summit, one of the questions posed to me and my team was our position on deregulation in the utilities sector, that is, allowing more competition into the market and enabling customers to invest in generating their own electricity, removing their reliance on the power grid. The deregulation movement began in the 1970s and as of 2018, 17 of the states in the United States have passed a form of deregulation into law.¹

I probed for the rationale behind the question and learned that they preferred to partner with those who were not in favor of deregulation, as it threatened their revenue and market leadership position.

My next question was whether they were investing in self‐generation systems and solutions so that they would be positioned to become the global leader if and when deregulation took place. I posited that if they were to develop and test such a product within markets and states that are not regulated, it could create a global revenue stream outside of their regulated footprint, providing ample testing ground to prove and refine the product.

This was not something they had considered. Rather than invest in the possibility of being a global leader in self‐generation of power, they were focused on ensuring that a change in the regulatory landscape did not take place, as it would present an existential threat to their organization. This is because they were focused on maintaining the existing system.

Centuries‐Old Systems

In the process of manufacturing, which could be considered the beating heart of the Industrial Revolution, a single anomaly can create a disaster. It might be something entering a machine that does not belong that could cause a machine to break down or threaten the safety of those working near the machine, or a slight damage to a component of a machine that, unchecked, produces hundreds of flawed products that need to be scrapped. There are both physical and fiscal reasons to streamline anything that can possibly be streamlined to avoid an anomaly.

Throughout our contemporary systems lay evidence that the Industrial Revolution effectively built the backbone of society in the late nineteenth century and throughout the twentieth century, evangelizing the effectiveness of streamlining and the evils of anomalies. Schools that have been standardized to one form of learning, standardized tests to measure intelligence or knowledge, standard operating procedures across industries, even the eight‐hour workday and the concept of a weekend away from work were invented during this era (the last two thanks to Henry Ford). In the contemporary era, it is difficult to pass a single hour of time without interacting with centuries‐old systems.

These systems were built on the foundation of viewing the world as a machine, each piece a subsystem of the bigger machine. The fallout of this worldview is that it requires people to operate as if they are components within a broader machine, leading them to feel like cogs.

As it stands in the early twenty‐first century, the world has been architected, with few exceptions, within the context of this mechanistic worldview. Whether an organization was built 100 years ago to create a new system or 30 years ago to reinforce a system, many organizations find themselves in an endless cycle of making improvements (reformation) but very few reimagine the original design. This is a natural outcome of the life cycle of organizations.

The majority of systems naturally reach a developmental stage where internal and external factors begin to pressure the system to maintain stability, and only change in small increments. This leads organizations to become trapped by the reliance their customers have built on their products and price points, forcing them into “maintenance mode.”

Maintenance mode is an organizational state in which preservation and incremental improvement become the central orientation of the organization.

There are three telltale signs by which an organization can be assessed for maintenance mode:

1. Criteria for initiatives across the organization include an orientation toward making incremental improvements without disrupting the system.
2. Approved initiatives improve individual functions without impacting broader organizational performance.
3. Organizational risk tolerance is extremely low, often indicated by a rigid return on investment justification requirement.

Directly following the release of a product or service and broad market adoption, signs 1 and 3 would be important indicators of long‐term continuity of the product or service. The same is true for operational leaders who need to maintain continuity of the organization's core value proposition. These are ideal arenas for reformation; although transformation can and does apply to these organizations, the allowable risk tolerance is often better suited to reformational initiatives.

Where maintenance mode becomes a problem is its prevalence in an organizational culture across all functions and at all levels, and it is a natural outcome of much of the most popular management thinking and leadership frameworks of the twentieth century and early twenty‐first century.

Maintenance Mode as a Means of Sustaining Scaffolded Systems

Scaffolded systems are the most challenging to change. The scaffolding of systems that comprise the utilities sector, for example, contains a subsystem of power generation, which contains further subsystems of coal refineries, nuclear generators, solar farms, and wind farms. The utilities sector also contains a subsystem called the energy grid, which, oversimplified, delivers energy to everything from streetlamps to skyscrapers within its geography. The utilities sector, examined as a part of the larger system of society, serves the critical function of supporting medical facilities, data centers, computers, and communication networks. A deeper layer to this complexity is that utilities, in developed countries, is a heavily regulated sector, meaning there are legal policies in place governing activities such as how utilities organizations spend their money, how they communicate with their customers, and where and how they store their data.

As you can imagine, introducing a change to this scaffolding of systems is difficult not only from a complexity point of view, but also from the vertical scaffolding point of view, in that the implications in the case of a system failure are disastrous.

Within the context of a scaffolding of systems such as these, investment in initiatives focused on emerging technologies is understandably met with systemic resistance.

If a new island with a landmass equal to the size of Europe appeared in the Pacific Ocean and its utilities sector could be designed from the beginning (as an act of creation), with all the technologies currently available and without existing utilities systems in place to maintain and keep it stable, it would undoubtedly be the most advanced utilities system on the planet. The path from an existing system to a more advanced system or set of scaffolded systems, however, is much more challenging and counterintuitive than the incremental, return on investment–based accounting in place within the majority of organizations today.

Each industry vertical is an interconnected network of scaffolded systems that faces three primary impediments to system‐wide change.

First, the complexity of scaffolded systems makes the consideration of change an exercise that involves a higher degree of complexity, vulnerability, and risk than any other business activity in empathy, research, ideation, analysis and synthesis, prototyping, testing, partnership, negotiation, leadership, and so on.

Second, in most organizations, no one below the executive office is accountable, in terms of measures tied to incentives, for improving the performance of the system as a whole or for improving the interaction of their part of the system with the rest of the system. The leaders of each part are incented and relied on to ensure that their part of the system continues to operate smoothly and at the highest possible performance. At its worst, when an initiative fails or organizational performance degrades, this orientation passes the blame from part to part, or team to team, when the failure is actually a product of the interactions within the whole system and not an individual part or team.

Third, accounting systems within organizations account for actions and risks taken and whether or not they were successful, but not for actions and risks that were not taken and whether or not they would have been successful.

Maintenance Mode and Advanced Technologies

Against this backdrop, there is an indication as to why there has been little meaningful adoption of artificial intelligence and other emergent technologies. Organizations in maintenance mode naturally reject anything that would introduce a system‐wide change and therefore threaten the stability of the system on which its customers rely. Pressured by shareholders, leaders, or customers, these organizations have a natural tendency to explore emergent technological advancements through pilots, a situation in which rigorous justification of the pilot was not required, and therefore there is no path to production. For many years, technology and advisory firms engaged happily in pilots, often offering investment eager to demonstrate their capabilities. In more recent years, however, as pilot after pilot has failed to convert to a larger partnership regardless of its ability to achieve the goal of the pilot, this model has begun to be called into question by technology leaders. This prevalent phenomenon has been termed “pilot purgatory.”

Other organizations in maintenance mode have responded to these technologies by hiring or promoting a leader to build a team or capability around these technologies and partner with the rest of the organization to identify, design, and execute initiatives to achieve efficiencies. For many organizations, this model has yielded better results than the pilot purgatory model, but not without significant investment and internal resistance. Given the small number of data and decision scientists in the market, combined with organizational leaders who may not have the expertise to accurately assess the trustworthiness and credibility of these experts, many of these budding teams spend years researching without a clear path to implementation or return on investment and are subsequently shut down or face shuffles of leadership in hopes of achieving the right organizational formula to generate practical value.

This is further complicated by a lack of practical experience and the orientation of many who go into the field of data science, who are trained and recognized for seeking breakthrough applications to extend the scientific field, leading to patents, research papers, and awards. If data scientists are particularly enterprising and emerge with a doctor of philosophy (PhD) at the age of 25 and enter the market, they have approximately 40 working years, or 80 projects (given a generous assumption of only six months per project throughout their career) in which to leave their mark on the world of science and academia. This context, combined with the scarcity of data scientists and the resultant supply of high‐paying options, renders it difficult to hire, train, and retain data scientists to solve known or mundane organizational needs, regardless of their monetary value. The organizational desire to leverage data science, which remains a nascent field, to make low‐risk, incremental improvements to the system is at odds with the interests and intrinsic motivation of many data scientists.

An analogous equivalent to this dilemma would be if corporations were to hire budding astronauts and ground them, as space presents too much risk, and instead wanted them to recreate research findings achieved by other astronauts who did go to space, translating their findings into projects specific to their organization to achieve a decrease in cost or increase in profit.

Maintenance mode is at odds with exploration, and organizations in maintenance mode that endeavor, with the best of intentions and even with significant investment, will struggle to harness the potential of artificial intelligence and its adjacent technologies.

Is Technology the Problem or the Solution?

The lack of successful implementations of advanced technologies has led to a self‐reinforcing cycle of negative outcomes, leading to increasing skepticism and the assignment of blame on the technology or on those implementing the technology. This is difficult to ratify with the incredible feats technology organizations have been able to achieve, subsequently making up 7 of the 10 most valuable public companies in the world at the time of this writing, and five of the seven companies in the world whose market capitalizations have, at any point, exceeded $1 trillion.

The technology is neither the problem nor the solution alone. Technology organizations that have leveraged advanced technologies within the context of acts of creation in the market have transformed markets and achieved new heights of value creation. In contrast, technology organizations, arguably with the same degree of talent and, in many cases, more assets at the onset, that have not performed acts of creation or transformation in recent history but have been stuck in maintenance mode, have not been able to grow by the same order of magnitude as technology organizations that either created from nothing (such as Google, Netflix, and Amazon), or managed to overcome maintenance mode and perform acts of creation and transformation (such as Apple and Microsoft).

Even at the executive level, there can be a disconnect between incentives and accountability tied to system maintenance as opposed to acts of creation and transformation. In publicly traded companies, for example, a large portion of executive incentives are made up of shares of the corporation, the value of which is not tied to the performance of the system in its function as a part of a larger system, but to the perception of shareholders, largely based on sustained growth. Acts of creation, which require strategic, multiyear investment, can negatively impact financial performance of the organization in the short term, leading to decreased shareholder confidence and a decrease in pay for the leadership team and any employees participating in profit‐sharing or stock option programs. Maintenance mode, although unsustainable in the long term, is profitable in the short term.

This prevalence of maintenance mode and its related metrics can be observed in the naming convention of the event that takes place each quarter that has the biggest consistent impact on stock trajectory: the earnings call. It is not a performance call, although the terms are often used interchangeably. It is also not a strategy call. If the function of an organization in a given subsection of the market is to provide necessary goods for the home at an affordable price, and the leadership announced that, in a given quarter, they had managed to improve the system so much that they could drastically reduce prices and maintain the same profit margin, the performance of the company in its essential function within society would have been improved, but unfortunately, shareholder confidence in the leadership team would likely decline along with its stock performance.

It is exceedingly difficult to introduce system‐wide change from within a system. Out of the five well‐known, system‐wide changes in the following table, only one of them was introduced from within the system. American Airlines introduced a technology in 1978 to enable travel agents to make reservations using an online system, which was ultimately leveraged to create Travelocity 18 years later. Otherwise, the system‐wide changes, often despite the fact that the change was initially conceived from within a market leader within that industry vertical (such as within Kodak and Blockbuster), were introduced from outside these organizations—not because these organizations tried and failed, either. The conception never gained enough internal traction within the large organizations to create a new line of business or introduce a product or service that could disrupt the whole system.

The Coexistence of Maintenance and Creation

The process of overcoming maintenance mode is not a cross‐organizational pivot away from maintaining the systems on which customers rely to the act of creating new products and services in the marketplace. Rather, some resources should be devoted to maintaining system stability while other resources pivot to creating a vision for the future of the organization within society, the market, and its place in time. This vision should not take into account the current system during the vision‐setting phase. Rather than beginning the process with the question “Where do we go from here?,” the process should begin with “What ought to be?”

Answering the question of “Where do we go from here?” can lead to fine‐tuning ships and inventing new tools for whaling in 1890, more aerodynamic carriages in 1908, or analyzing video footage of Blockbuster stores to create heat maps of foot traffic in 2007.

Once the question of “What ought to be?” is answered, the next question is “How do we get there from here?”

This process will be discussed further throughout this book.

A Note for Individual Contributors

One of the most impactful conversations I had during my tenure at Microsoft was with my former boss, Jennifer Byrne, when she was the chief technology officer for Microsoft US. I was expressing frustration over the metrics that had been tied to a portion of the budget our team had received from another team, as I felt that achieving those metrics would require operational focus on those numbers disconnected from and even opposed to the potential for market impact that had attracted me to the position.

“Screw the metrics,” she told me.

She went on to ask me what market impact had attracted me to the position, and what I wanted to achieve within the next year. (“What ought to be?”)

I answered her question, and she asked me: “If you achieved those things, wouldn't you blow the metrics out of the water?”

Even in a position where I did not have full control over all the metrics assigned to me or tied to budgetary allocation, this pivot enabled me to focus on the impact I wanted to have and see the metrics as milestones on that journey, as opposed to chasing the metrics for metrics’ sake, which could have been achieved without creating lasting market impact.

The same holds true in an individual contributor position: if you do not have control over your metrics, the first step is to imagine the market impact you would like to create, and see the metrics as milestones on that journey. If the market impact you imagine is incompatible with the metrics you have been assigned, you can summarize the market impact you would like to achieve together with proposed new metrics, and if that does not work, it might not be possible to create the market impact you want to create at this stage in your career within the system you are in, and it might be time to look for your next adventure either internal or external to the organization.