The Zone of Control: The Gateway to Operational Excellence

Imagine a U.S. metal‑stamping plant that installs an AI‑driven scheduling system to boost throughput. Instead of efficiency gains, work‑in‑progress inventory balloons, and output per labor hour falls. The new system's recommendations clash with undocumented changeover routines on the shop floor, creating bottlenecks at every press. This kind of story is illustrative, not unique. Large‑scale evidence shows that when firms automate before stabilizing their processes, technology often magnifies flaws rather than fixes them. 

A recent U.S. Census Bureau study of manufacturers found that companies adopting industrial AI frequently experience an initial decline in productivity before later gains, precisely because automation collides with fragmented workflows and inconsistent practices. In other words, technology doesn't cure chaos; it accelerates it until organizations establish control.

This story is not unique. Hospitals that deploy AI‑driven diagnostic tools without standardized care pathways see little improvement in patient outcomes. Utilities that invest in predictive maintenance software without unified work order systems continue to suffer outages. Telecom firms that launch innovation labs without consistent reporting still struggle with reliability.

The lesson is not that technology fails, but that sequence matters. Research shows that digital transformations stumble for many reasons, such as cultural resistance, poor leadership alignment, or lack of skills, but one of the most overlooked is the absence of control. Until work is standardized, visible, governed, and measurable, improvement is an illusion.

Threshold 1: Getting organized

Every executive has seen the symptoms of an organization that has yet to cross the first threshold. Employees spend late nights juggling spreadsheets, copying data from one system into another. Managers constantly firefight, rushing to fix errors, soothe frustrated customers, and patch over gaps in reporting. Success depends on a handful of "heroes" who know how to navigate the chaos, but when those individuals leave, performance collapses.

This is the stage of managed disorder. Work gets done, but at enormous human cost. Customers experience delays and inconsistencies. Employees burn out. Leaders are left with anecdotes instead of evidence.

Crossing the first threshold means moving from improvisation to consistency. Processes are documented, responsibilities are clarified, and basic workflows are followed. In manufacturing, this might mean standardizing assembly‑line procedures so that every shift operates the same way. In healthcare, it could mean codifying care pathways so that patients with similar conditions receive consistent treatment. In telecom, it might mean creating a single playbook for outage response rather than relying on the instincts of individual managers.

The transformation: Once organizations are organized, they gain their first layer of stability. Errors decline because people are no longer reinventing the wheel. Employees feel less burned out because they are not constantly improvising. Leaders gain their first glimpse of patterns in performance, even if the data is still limited. This stage creates the foundation for trust. Employees know what is expected, and leaders know that work is being done consistently.

The leadership imperative: At this stage, leaders must insist on documenting processes, even when teams resist. Standardization can feel bureaucratic, but without it, no further progress is possible. Leaders who frame organization as a way to free people from chaos, not as red tape, are the ones who succeed.

It is tempting to deploy AI or automation at this stage, but doing so is premature. Without consistent processes, technology simply accelerates inconsistency. A chatbot trained on fragmented customer data will frustrate users with contradictory answers. A predictive model fed with inconsistent production data will generate unreliable forecasts. At this threshold, the role of technology is modest: simple workflow tools, shared dashboards, and basic reporting systems that reinforce consistency. The real work is cultural, teaching people to follow the same playbook.

Threshold 2: Gaining control

If the first threshold is about reducing chaos, the second is about establishing authority over the system itself. This is the moment when an organization stops lurching from one fire drill to the next and begins to operate with discipline. It is also the most difficult threshold to cross, because it requires leaders to make unpopular choices: mandating common platforms, enforcing data standards, and dismantling cherished local practices.

At this stage, the organization's challenge is not a lack of effort but a lack of coherence. Different business units may each have their own systems, their own definitions of success, and their own ways of reporting results. Leaders may spend more time debating whose numbers are correct than deciding what to do about them. Employees, meanwhile, are caught in the middle forced to reconcile conflicting instructions and duplicate their work across multiple systems.

Crossing the threshold into control requires integration. In a utility, it might mean unifying work order systems so that field crews, schedulers, and regulators are all working from the same data. In healthcare, it could mean consolidating reporting across departments so that adverse events are tracked systemically rather than as isolated incidents. In manufacturing, it often means connecting production data, quality metrics, and supply chain information into a single view.

The transformation: Once control is established, leaders can finally trust their data. They can distinguish between normal variation and true anomalies. They can enforce rules that everyone follows, ensuring that processes are applied consistently across the enterprise. Decision‑making accelerates because leaders are no longer debating whose numbers are correct; they are acting on a single line of sight across the business.

This is also the point at which advanced technologies, including artificial intelligence, begin to deliver real value. Predictive models can only function when the data feeding them is standardized and reliable. Without control, AI amplifies noise; with control, it amplifies insight. A machine‑learning model trained on inconsistent data will produce inconsistent recommendations. But once control is in place, the same model can reveal systemic risks, optimize resource allocation, and surface opportunities that humans might miss.

Consider the case of a North American electric utility that had invested heavily in predictive maintenance software. Despite the investment, outage rates continued to rise. The problem was not the algorithms but the inputs: work orders were coded differently in each region, and reporting systems were fragmented. Crews were dispatched based on incomplete data, and leadership lacked visibility into systemic issues. Only after the company mandated a single work management platform, standardized workflows, and enforced governance did the predictive tools begin to deliver value. Within 18 months, outage response times fell by 22 percent, and regulatory penalties declined sharply.

The lesson is clear: control is not about bureaucracy for its own sake. It is about creating the conditions under which technology can succeed. Leaders who fail to enforce control will find that their digital investments underperform. Leaders who cross this threshold discover that the very same tools suddenly become powerful multipliers of performance.

Threshold 3: Unlocking excellence

The final threshold is where organizations move from control to optimization. With trustworthy data and enforced discipline, they can now systematically apply advanced analytics, automation, and innovation.

In utilities, predictive maintenance algorithms can reduce outages because they are fed with consistent, high‑quality data. In healthcare, AI‑driven diagnostic tools can flag early warning signs of sepsis because care pathways are standardized and outcomes are comparable. In manufacturing, digital twins and machine learning models can optimize throughput and reduce defects because the underlying processes are stable.

The transformation: At this stage, improvement becomes self‑reinforcing. Organizations shift from reactive firefighting to proactive problem‑solving. They can align resources with enterprise priorities, reduce variability, and accelerate innovation. Technology becomes a multiplier of disciplined processes, not a substitute for them. Leaders who once struggled to extract value from digital investments now find that those same tools deliver exponential returns.

The cultural shift is equally significant. Employees move from compliance to engagement because they can see how their work contributes to enterprise goals. Continuous improvement becomes part of the organizational DNA. Innovation is no longer episodic or dependent on a few visionaries; it becomes systematic.

Consider a large urban hospital that initially invested in AI‑driven monitoring devices and diagnostic tools. Outcomes barely improved until leadership standardized care pathways across departments and integrated reporting into a single dashboard. Once processes were consistent and data comparable, the hospital could identify systemic risks rather than isolated incidents. Within two years, adverse events declined by 30 percent, echoing findings from the Journal of Patient Safety. Only then did AI tools become powerful, flagging early warning signs of sepsis and highlighting systemic risks across departments.

A manufacturing example illustrates the same principle. A global automotive supplier invested in robotic automation to improve throughput. Instead, defects multiplied, and robots simply produced flawed components faster. The breakthrough came when the company standardized quality checks and integrated production data across plants. Once control was established, automation and AI‑driven digital twins began to optimize throughput and reduce defects. The same technology that once magnified problems became a driver of excellence.

The leadership imperative: At this stage, leaders must tie analytics and AI to strategy, not novelty. Otherwise, organizations risk "innovation theater"; flashy pilots that don't move enterprise outcomes. Leaders who succeed are those who treat governance not as red tape but as a strategic asset. They understand that discipline is not the enemy of innovation; it is its precondition.

Earning the right to optimize

Operational excellence is not achieved by skipping ahead to optimization or innovation. It begins with the discipline of control. The threshold for control is the most significant milestone in the journey, as it is only after crossing it that organizations can truly improve.

The question for executives is not whether to optimize, but whether they have earned the right to optimize. Until the threshold is crossed, improvement is an illusion. Once crossed, it becomes inevitable.

Three questions every CEO should ask before optimizing

1. Can I trust the data I'm seeing? If different teams are working from different numbers, optimization efforts will be wasted.
2. Are our processes consistent across the enterprise? If every department or plant has its own way of working, technology will only magnify inconsistency.
3. Do we have rules that everyone follows?  Without accountability for following standard processes, even the best tools and analytics will fail.

References and supporting evidence

• Ghosh, S., Hughes, M., Hodgkinson, I., & Hughes, P. (2022). Digital transformation of industrial businesses: A dynamic capability approach. Technovation, 113, 102414.
• Gfrerer, A., Hutter, K., Füller, J., & Ströhle, T. (2020). Ready or not: managers' and employees' different perceptions of digital readiness. California Management Review, 63(2), 23‑48.
• Journal of Patient Safety. "Impact of Standardized Care Pathways on Adverse Events," Vol. 17, No. 4, 2021.
• McElheran, K., Yang, M., Kroff, Z., & Brynjolfsson, E. (2025). The Rise of Industrial AI in America: Microfoundations of the Productivity J-Curve(s). Presented at CES, April 25–27.
• MedRxiv. The use of Artificial Intelligence in the out‑of‑hospital care settings: A systematic review, 2025.
• MIT Sloan Management Review. A Maintenance Revolution: Reducing Downtime With AI Tools, 2023.
• North American Electric Reliability Corporation (NERC). Human Performance and Root Cause Analysis Reports, 2022.
• OECD Observatory of Public Sector Innovation. The Devious Dozen: 12 fallacies behind innovation lab "failures", 2022.
• Pumps & Systems. AI‑Driven Asset Failure Prediction in Water Utilities, 2024.
• U.S. Government Accountability Office (GAO). High‑Risk Series: Federal IT Modernization, 2023.
• Turrin, R. 12 Reasons Why Most Innovation Labs Are Failing, 2021.