Operating Model for a Manufacturing Enterprise Knowledge Document

Challenges for the Adaptability of the Operating Model

Modern manufacturing operating models were largely designed for a different era—one with slower markets, longer product cycles, more stable workforces, and clearer functional boundaries. Those assumptions no longer hold. As a result, many operating models struggle to adapt under current conditions. The most visible challenge is decision and permission latency. Organizations often spend months aligning, socializing, and seeking approval for actions they already know are necessary. By the time execution begins, the window in which the improvement would have mattered most has narrowed or closed entirely. This delay is rarely caused by a lack of effort; it is caused by architectures that treat decisions as events rather than as outputs of a system. Workforce dynamics amplify the problem. Shorter tenure and higher churn mean that operating models increasingly rely on expertise that no longer exists at scale. Tacit knowledge walks out the door faster than it can be codified, while remaining employees are asked to manage greater complexity with less experience. At the same time, asset and product complexity continue to rise. Expanded product portfolios, more sophisticated equipment, and tighter regulatory constraints increase the cognitive and coordination load placed on the organization. Traditional hierarchical models respond by adding layers of review and control, which further slows response times. Governance, while essential, often becomes another source of friction. Capital allocation, safety, quality, and compliance processes are frequently optimized for risk avoidance rather than value creation. When governance cannot operate at the same speed as the operating model, it becomes the bottleneck. Finally, technology adoption has outpaced organizational redesign. Advanced analytics and AI are layered onto operating models that were never designed to act at machine speed. The result is more insight, more dashboards, and more alerts—without a corresponding increase in action.

Orchestration AI within the Operating Model Framework

Orchestration AI should not be understood as a technology initiative. It is an operating capability. Its value is realized only when it is embedded into how work is prioritized, decisions are made, and actions are executed. In effective operating models, AI does not sit on top of the organization in the form of dashboards or reports. Instead, it lives inside workflows, management routines, and decision rights. Its primary role is not to generate insight, but to orchestrate attention, recommend actions, and help execute decisions within clearly defined guardrails. A critical distinction is between correlation and causality. In highconsequence or tightly governed decisions, AI that can support interventiongrade reasoning materially improves trust and permission. However, many operational gains can still be achieved through correlationbased models when actions are preauthorized within clear policies and safe operating envelopes. The operating model must distinguish between where causal reasoning is required and where bounded inference is sufficient. Orchestration AI must support intervention—helping leaders understand what will happen if they act, not just what has happened before. This causal capability is what allows AI-generated recommendations to earn trust and permission. Equally important is respect for governance. Orchestration AI must be permission-aware, operating within safe operating envelopes defined by safety, quality, and compliance requirements. When designed correctly, it reduces latency by making boundaries explicit rather than by bypassing them. The end state is not autonomy for its own sake. It is AI-in-the-workflow: systems that scale expert judgment, reduce cognitive load, and accelerate execution while keeping humans accountable for outcomes.

Adopting Fails to Accelerate Value: A Perspective

Many organizations adopt new tools enthusiastically yet see little acceleration in value. The root cause is not technology failure, but operating model inertia. Adoption is treated as deployment rather than as a change in how decisions are made and executed. When operating rules remain unchanged, new tools simply surface more insight into an unchanged system. Dashboards proliferate. Pilots multiply. Activity increases, but outcomes do not. Leaders mistake motion for progress and declare success while productivity remains flat. True value acceleration begins earlier than most organizations expect. It begins when decision time and permission time are treated as part of execution rather than as overhead. It continues when capital is allocated in bounded stages that deliberately fund learning instead of demanding certainty upfront. And it requires that decision rights, governance mechanisms, and management routines evolve alongside technology. Without these changes, adoption improves visibility but not performance. The organization looks more advanced while behaving the same way it always has.

Adapting Operating Models to Fit Your Reality

There is no universal best operating model. Effective models are designed around constraints, not aspirations. They reflect the realities of the market, the nature of demand, the variability of supply, and the limits of the workforce. Some organizations are market-constrained, struggling to respond quickly enough to changing customer signals. Others are production-constrained, limited by asset capacity or reliability. In each case, the operating model must focus on the decisions that matter most under those conditions. Adaptation is most successful when it begins with a small number of high-impact decisions rather than with enterprise-wide redesign. Decision rights should align with proximity to information, while governance should be designed to bound risk without paralyzing action. Over time, the operating model must be allowed to evolve through use, feedback, and learning. The strongest operating models are not the most elaborate. They are the ones that are simple enough to execute consistently and robust enough to adapt as conditions change.

Operating Models Across Industry Sectors

Operating models tend to reflect the dominant economic and technical realities of their industries. In automotive, lean and Toyota Production System–derived models dominate. These emphasize standardization, takt-driven flow, and deep supplier integration. They excel at cost, quality, and repeatability, but increasingly struggle with differentiation, software-driven complexity, and the tension between speed and governance. Aerospace and defense organizations typically operate through program-centric models with heavy compliance and stage-gate governance. These models prioritize safety, reliability, and traceability, but often suffer from extreme decision latency and difficulty scaling AI under regulatory constraints. Chemical companies tend toward asset-centric, reliability-driven operating models. Continuous processes, capital intensity, and safety considerations dominate. These models deliver strong yields and utilization, yet face challenges managing portfolio complexity and sustaining disciplined capital renewal. Food and beverage organizations are often demand-driven, with operating models optimized for service levels, freshness, and responsiveness. High SKU counts and short cycles reward speed, but create labor intensity and margin pressure that strain traditional management systems. Life sciences organizations—particularly pharmaceuticals, biotech, and advanced therapies—operate under some of the most constrained operating models in manufacturing. These models are shaped by extreme regulatory oversight, long development timelines, and an uncompromising emphasis on patient safety. Decision rights are tightly coupled to validation, documentation, and compliance, often resulting in highly formalized stage gates and approval structures. While these operating models excel at traceability, quality, and risk mitigation, they frequently struggle with speed, scalability, and cost. The rise of personalized medicine, cell and gene therapies, and rapiddemand therapies has exposed the limits of traditional batchcentric and inventoryheavy models. In these advanced therapy contexts, inventory may be minimal or lose practical meaning altogether, unlike in traditional pharmaceutical manufacturing where inventory remains a core planning lever. In many cases, inventory either cannot exist or loses meaning entirely, forcing organizations to rethink planning, scheduling, and capacity management from first principles. For life sciences, the operating model challenge is not whether governance is necessary—it is how governance can be redesigned to enable faster learning and decision-making without compromising compliance. Emerging models increasingly embed digital twins, advanced analytics, and AI directly into validated workflows, allowing organizations to close the gap between signal and response while maintaining accountability. The differentiator is no longer compliance alone, but the ability to execute compliant operations at speed.

Cultural Implications of Operating Models

Every operating model encodes a culture, whether intentionally or not. Hierarchy-heavy models reinforce risk aversion and consensus-seeking behavior. Distributed models require trust, clarity, and a high tolerance for accountability. As AI becomes embedded into operating models, cultural implications intensify. People must feel safe acting on machine-generated recommendations, while leaders must be willing to accept bounded risk in exchange for speed. Where fear dominates, permission expands and velocity collapses. Culture ultimately determines whether signals lead to action or to meetings, whether failure produces learning or blame, and whether operating models evolve or calcify over time.

Elements of a Typical Manufacturing Operating Model

A manufacturing operating model is best understood as a system rather than a set of components. At its core is a decision architecture that defines who decides what, when, and with what evidence. Around it sit governance mechanisms for capital, safety, quality, and risk, designed to enable speed within boundaries rather than to enforce consensus. Management systems translate decisions into daily, weekly, and monthly routines. Operational excellence provides the discipline that makes those routines repeatable and reliable. It defines the standards, management behaviors, and continuous improvement mechanisms that stabilize performance while creating capacity for change. Without operational excellence, operating models oscillate between heroics and initiatives rather than delivering sustained results. Knowledge orchestration is the connective tissue that allows operating models to scale beyond individual expertise. It extends traditional knowledge management and standard work by ensuring that contextualized knowledge—lessons learned, expert judgment, validated practices, and operational constraints—is actively delivered to the point of decision. Unlike static repositories or documentation systems, knowledge orchestration is embedded in workflows and decision processes, enabling consistent execution without requiring consensus or reinvention. It governs how institutional knowledge is captured, contextualized, and reused across roles, sites, and time. In many organizations, knowledge exists but is fragmented—trapped in people, documents, dashboards, or disconnected systems. Knowledge orchestration ensures that insights, lessons learned, and expert judgment are delivered to the point of decision, when they are needed, in a form that can be acted upon. Technology then integrates IT, OT, and operations around decision flow rather than data flow alone, enabling both operational excellence and knowledge orchestration to function at scale. Performance feedback loops reinforce learning by highlighting what matters, not by overwhelming the organization with noise. When these elements work together, the operating model becomes a compounding advantage. When they do not, even the best strategy struggles to survive contact with reality.

This document reflects LNS Research perspectives and COO Council discussions on operating models, productivity, and the role of Industrial AI in modern manufacturing.

Operational Excellence Is Not Your Operating Model

James Wells Jul 30, 2025 Pathfinder companies in the LNS Research Industrial Productivity Index™ that are outpacing peer companies’ long-term industrial productivity growth have several areas where their practices differ from those of the general index companies. One of those areas is an Operating Model. We see companies that think about their Operational Excellence approach as their operating model. It is easy to conflate Operational Excellence (OpEx) with an Operating Model. After all, both aim to improve performance, reduce waste, and ensure sustainable results. Understanding the difference is critical for organizations aiming to evolve beyond legacy improvement practices toward integrated, agile, and resilient industrial transformation. What is an Operating Model? We define an Operating Model as: A principles-based approach that defines the way of work at a company that is aligned with: • • • Values • Goals • Strengths • Strategies An Operating Model (Figure 1) is not a single function or set of practices. It is a framework that defines how work gets done at a company. It spans five (or more) interdependent elements: • • • Leadership • Culture • Respect for People • Value Generation Principles • Operational Excellence Figure 1: LNS Research Operating Model Framework Operating models can and should vary by company, but Operational Excellence plays a crucial role in all of them and is one of several foundational components. While Op Ex typically focuses on performance management, waste reduction, continuous improvement (CI), and sometimes risk, the Operating Model defines everything, including inventory flow, product lifecycle processes, governance structures, talent systems, digital architectures, risk frameworks, and value propositions to the customer. Operating Models define the ethos and guiding principles that a company subscribes to. Adapt or Adopt? Many industrial companies have adopted the latest fad in the news in hopes of finding that silver bullet that will align their efforts and produce significant business value. We have seen it in Operational Excellence in the past. GE had a good run with Six Sigma in the late 80s and early 90s, and Toyota made news for its lean efforts, so seemingly everyone hopped on the bandwagon, hoping for similar newsworthy results. Similarly, for Operating Models. Our research shows that the motion many of these companies have taken is to start with adopting a known model’s core principles and values…but then adapting the tactics and approaches to the company’s own unique situation and value proposition. We have seen this with the Mercedes-Benz Production System (Figure 2), Danaher Business System, and General Electric Aviation’s Flight Deck, all adapted from the Toyota Production System; Illinois Tool Works 80/20 Operating Model, adapted from the simple Pareto Principle; and Tata’s Business Excellence Model, adapted from the Malcolm Baldrige Quality criteria. Figure 2: Mercedes-Benz Adapted Operating Model Some things that we see across these examples and several others are: • • • Brand it. • Use it as the framework for Change Management within the company to drive buy-in. • Create senior executive roles to nurture and own the model. • Push up and down the value chain. • Digitize critical elements of it to make it more sustainable. The Operating Model Demands a Wider View Operating Models typically integrate: • • • Embedded Quality as well as Integrated Operational Excellence • Future of Industrial Work, Knowledge Management, and Change Management • Embedded EHS and Intelligent Risk Management • Intelligent Supply Networks and Resilient Architecture • Critical players on the value chain, including suppliers, engineering, planning, delivery, inventory, manufacturing, and service functions Operational Excellence alone cannot accomplish this convergence; it requires a unified view — architected intentionally — where OpEx is the improvement engine within a broader system of value creation and brand protection. Operational Excellence: A Powerful Enabler, Not the Whole Structure Traditional Operational Excellence has been rooted in the CI and waste reduction practices of Lean, Six Sigma, TPM, or world-class manufacturing, which are primarily procedural approaches. While valuable, these methods have relied heavily on expert knowledge, high degrees of operational discipline, manual execution, and long project cycle times. These traditional CI systems face existential headwinds: talent turnover, slowing results, and limited scalability. In our research on Operational Excellence in 2023, 75% of companies reported that their traditional approaches were losing momentum (Figure 3). Enter Integrated Operational Excellence — a digitally enabled evolution of operational excellence. Integrated OpEx embeds OpEx tools and practices within a digitally enabled operating environment, leveraging tools like Digital Twins, AI/ML, and Advanced Analytics to transform how work is improved. This shifts Operational Excellence from a standalone, siloed initiative to a functionally integrated and dynamic element of a broader Operating Model​. Implications: Rethink How You Deploy Op Ex To thrive in this new paradigm, organizations are repositioning Operational Excellence in three key ways: • From Paper to Digital. Rather than viewing Op Ex as a static program with Lean toolkits implemented by experts, elevate it to a digital platform capability within the Operating Model — one that connects people, data, and machines in real time. • From Siloed to Integrated. The fast pace of modern manufacturing demands that solutions keep up. Traditional, asynchronous approaches to Operational Excellence are severely outdistanced by digital. Use digital tools (Digital Twins, AI/ML, Analytics) to monitor, model, and manage performance in near real-time. According to our research on Operational Excellence, leaders are 7 times more likely to connect analytics, AI, and Op Ex execution with the goal of faster, better decision intelligence. • From Procedural to Intelligent and Strategic. Align Operational Excellence with other pillars of the Operating Model. For example, embedding Intelligent Risk Management within Digital Operational Excellence enables organizations to monitor and respond to disruptions impacting operational agility, flexibility, and resiliency, potentially avoiding operational disruptions. Integrating Op Ex into the Operating Model unlocks talent development channels and innovative ways of delivering value to the customer, among other benefits. Figure 4: Op Ex is a Strategic Pillar Recommendations: A Structural Shift, Not a Semantic One Operational Excellence as a part of, but not the entirety of your Operating Model, is not a matter of semantics — it’s a structural shift. OpEx is essential, but it cannot address the full breadth of the complexities of the way of work at industrial companies today. To meet the demands of complexity, speed, and sustainability, organizations are embedding OpEx into a robust, adaptable Operating Model — one that defines how work is done, improved, and transformed across the entire enterprise. • Adopt an Operating Model that closely aligns with your core values, strategies, and ways of working. The only company that fits the Toyota Production System (TPS) is Toyota, but it or one of the others might be close to your unique way of working as a starting point. • Adapt that Model to your unique way of working. No model is a perfect fit for your company except for the one that your company customizes to your unique strengths, values, and strategies. • Integrate Digital Op Ex as a central pillar into the operational architecture to sustain long-term value from your efforts and align Op Ex efforts with business priorities and strategies. In short, Operational Excellence is the engine, but the Operating Model is the vehicle. To drive true business value, you need both, working together in harmony.

Pressure Test Assessment

Strengths vs. the market

COO-relevant framing (operating model as an execution system).
Your document is strongly aligned with the dominant “operating model as a system” view in major consulting and practitioner literature, not the narrower “org chart + processes” view. That matches how McKinsey, Bain, and Deloitte describe operating models: interlocking elements (governance, structure, processes, technology, behaviors) designed to deliver strategy. Decision rights / governance latency as a first-order constraint.
The emphasis on decision/permission friction is consistent with what BCG and others call out explicitly: unclear or contested decision rights delay decisions and erode confidence. That’s a credible, market-aligned backbone for a manufacturing operating model POV. IT–OT–Ops convergence positioned as an operating model issue, not just architecture.
Market guidance increasingly argues that convergence requires operating model changes (governance, roles, responsibilities) before or alongside technical transformations—your document matches that. Pragmatic stance on AI (human accountability and “in-the-workflow”).
Your cautious, “human accountability remains essential” posture is consistent with mainstream enterprise guidance on agentic/industrial AI, which emphasizes human-in-the-loop / oversight patterns. Life Sciences section is directionally correct on validation + governance.
Your claims about validation, documentation, and compliance shaping decision rights are consistent with FDA’s lifecycle validation framing and with ISPE’s GAMP emphasis on patient safety, quality, and data integrity through validated computerized systems. Weaknesses vs. the market (and where truthfulness is overstated) 1) “Decision latency is the primary constraint” can read as over-claiming.
In the market, productivity stagnation is typically treated as multi-causal (technology diffusion, capital allocation, management practices, competition dynamics, etc.). Your document is compelling as a POV, but it is not universally true that operating-model decision latency is the primary constraint relative to all other factors, across all manufacturers. If you present it as often or frequently the hidden constraint (rather than the constraint), it becomes more defensible. 2) “Causal over correlational intelligence” as a general requirement is too strong.
It’s directionally right that intervention-grade reasoning improves trust and permission, but the document implies (at points) that orchestration AI “must” be causal to create value. In practice, large value pools in manufacturing (e.g., predictive maintenance, anomaly detection, quality inspection) can be achieved without full causal identification—especially where actions are pre-authorized within bounded policies. Also, strong causal claims usually require assumptions that may be untestable and must be made explicit; otherwise the “causal” label becomes fragile. 3) Knowledge orchestration is a useful concept, but it’s not yet a market-standard term.
As written, it risks sounding like an invented layer unless you define it relative to widely recognized constructs (knowledge management, standard work, digital thread, data governance, semantic layer, etc.) and show how it differs. This is less a “falsehood” and more a positioning risk: the idea is right, the label may confuse buyers unless anchored. 4) Sector models are directionally correct, but they are high-level and can be challenged on specificity.
For example, “automotive = TPS derivatives” is broadly true, but many automotive groups now operate hybrid product-and-platform models (software-defined vehicle, centralized OTA governance, etc.). Similarly, “A&D = program-centric stage gates” is generally true, but leading programs are pushing more agile governance inside compliance boundaries. The statements aren’t false, but they are incomplete—your opponents will say “that’s yesterday’s archetype.” 5) Life sciences: “inventory can’t exist / loses meaning” needs narrowing.
That’s accurate for many patient-specific CGT chains (chain-of-identity, time constraints), but it’s not accurate across life sciences broadly (traditional pharma relies heavily on inventory strategies). If you narrow the claim explicitly to cell & gene / personalized therapies, it becomes materially more truthful and defensible. Truthfulness evaluation (overall) High confidence / well-supported (truthful as written): • Operating models are systems of interlocking design elements tied to strategy. • IT/OT convergence requires operating-model changes (governance, roles) alongside technical change. • Decision rights clarity reduces decision delay; unclear rights create latency. • In regulated life sciences, validated/controlled processes shape how fast change can happen. Partially true but needs qualifiers to avoid overreach: • Decision latency as the dominant limiter of productivity in manufacturing (should be “a major/frequent hidden constraint”). • “Causal AI” as a prerequisite for orchestration AI (should be “causal where interventions are high-consequence or permission gates require it; otherwise correlation + policy-bound action can still deliver value”). • “Life sciences inventory loses meaning” (true for many CGT scenarios; not generalizable to all life sciences manufacturing). Suggested edits to make it more market-proof (without changing your POV) • Replace “the constraint” language with “a frequent hidden constraint” in the decision-latency argument, and explicitly acknowledge macro drivers of productivity slowdown. • Add a short “when causality is required vs. when it isn’t” paragraph in the Orchestration AI section, grounded in the reality that causal claims require strong assumptions and may be bounded. • In Life Sciences, split into two sub-archetypes: Traditional pharma (batch/inventory/validation) vs Advanced therapies (CGT) (patient-specific, chain-of-identity, time-constrained). • Define “knowledge orchestration” by mapping it to market-recognized terms (standard work, knowledge management, digital thread, semantic context) and state what you mean operationally (who owns it, where it lives, how it changes decisions). If you want, I can implement these edits directly in the document while keeping your prose style and section structure intact.