Companies still run on a 20th century decision model

Executive summary

• Companies have digitized everything except how they make decisions.
• AI improves prediction, not alignment. Better inputs do not guarantee better outcomes.
• And increasingly, companies are realizing that better technology does not guarantee better outcomes either.
• The real bottleneck is structural and human: fragmented incentives, siloed organizations, misaligned objectives, and the absence of a shared language to arbitrate decisions across functions.
• What appears as a data or technology problem is, in practice, an orchestration problem.
• The next competitive advantage is the ability to organize decisions as a system driven by outcomes, not as isolated events.

‍For a long time, companies have tried to improve performance by improving the quality of individual decisions. It was a logical approach, and it delivered significant progress. But the next frontier lies elsewhere. The issue is no longer whether decisions are good in isolation. It is whether they are coherent together.

In the 20th century, organizations operated in relatively stable environments. Trade-offs were simpler, interdependencies were limited, and a degree of misalignment could go unnoticed. When conditions are stable, these inefficiencies tend to remain invisible. But in more volatile environments – shaped by geopolitical tensions, macroeconomic pressure, and continuous technological disruption – they become visible, and their consequences amplify across the organization.

Being technologically advanced does not make decision-making easier. It makes orchestration more critical.

‍

A silent mismatch

Spend time inside any large organization and a contradiction quickly appears. On one side, revenue-generating functions are highly engineered. Marketing models quantify ROI with increasing precision. Sales pipelines are tracked in real time. Pricing strategies are continuously optimized. Over the past two decades, companies have built strong capabilities to measure and improve commercial performance at scale.

On the other side, the moment decisions are made, the system becomes informal again. Choices are discussed in meetings, negotiated across silos, and based on partial representations of reality. Each function operates with its own metrics, constraints, and logic. Companies have modernized execution, but decision-making itself remains largely unstructured.

Decisions are still often shaped by experience, individual judgment, internal influence, or political dynamics, rather than by a shared, explicit logic of performance. Each function operates with its own metrics and priorities, and alignment happens late, if at all. These inefficiencies tend to remain invisible in stable environments. But when conditions become more volatile, they surface rapidly and propagate across the organization, amplifying risks and eroding performance.

‍

When good decisions don't add up

This is not a new problem. Herbert Simon described organizations as systems of "bounded rationality," where individuals make the best decisions they can within limited information and within the boundaries of their role. Russell Ackoff later formalized a related idea: optimizing parts of a system does not optimize the system itself. This also echoes the work of Daniel Kahneman on cognitive biases and decision-making under uncertainty, highlighting how even well-informed decisions are shaped by human heuristics, making consistent, system-level optimization difficult in practice.

Most companies still operate exactly this way. Each function is rational in isolation. A marketing team optimizes spend efficiency. A pricing team maximizes margin. A sales team pushes volume. Each uses increasingly sophisticated tools and data to make better decisions within its own scope. But those decisions interact. They reinforce or neutralize each other in ways that are rarely anticipated.

A classic example can be found in consumer goods. A marketing team increases promotional pressure to drive volume. Sales follows by pushing distribution. The result is short-term growth, but pricing power erodes and margins decline. Finance reacts by tightening budgets, reducing brand investment and weakening long-term demand. Each step is rational locally. The outcome is suboptimal globally. The system produces good decisions that do not add up to a good outcome.

‍

Why AI does not solve this

The current wave of AI, including agentic AI – like previous waves of data science, advanced analytics, or marketing mix modeling – is often presented as the solution to decision-making. In reality, it addresses a different layer of the problem. AI can be highly effective at prediction and optimization when the scope is clearly defined. It can improve forecasts, identify patterns, and enable faster, more granular decisions. But it does not resolve the fundamental issue of alignment across functions.

More accurate predictions do not guarantee impact if the chain from insight to decision to action is broken. The last mile – translating insights into consistent, coordinated decisions and execution – remains the hardest part. If anything, AI can intensify fragmentation and enhance silos. Each function becomes more capable, faster, and more confident in its own decisions, but not necessarily more aligned with others. The result is a paradox: more intelligence, but not more coherence.

In practice, capturing value from AI requires navigating a level of complexity that goes far beyond model performance. On the functional side, it means aligning decisions across marketing, pricing, and commercial strategy rather than optimizing them independently. On the data side, it requires structuring and connecting heterogeneous data sources into a shared, reliable foundation. On the technology side, it involves integrating models into existing systems and workflows, rather than layering tools on top of them. And on the organizational side, it requires driving adoption across teams, geographies, and functions, often in highly complex environments.

‍

From decisions to outcome-led systems

For years, most organizations have approached transformation through technology, assuming that better tools and models would naturally translate into better outcomes. In practice, this rarely holds.

If the issue is structural, the response cannot be incremental. What is missing is a system that organizes how decisions are made across the enterprise. One that creates a shared language of performance across functions, aligning how a CFO, a CMO, and commercial leaders define value, trade-offs, and success. It enables shared KPIs, shared taxonomies, and shared insights, breaking silos and resolving the contradictory signals that often exist between functions.

Such a system allows decisions to be evaluated in interaction rather than in isolation, makes trade-offs explicit before resources are committed, and ensures that execution follows through consistently. Over time, it turns decisions from discrete events into a coordinated process, orchestrated in service of business outcomes. This reflects a broader shift toward an outcome-led logic, where the starting point is no longer the technology itself, but the result to be achieved. In that sense, outcomes are not driven by technology alone, but by the coherence of the decisions that precede them.

‍

The orchestration problem

A useful analogy is that of an orchestra. Most organizations today are composed of highly skilled specialists. Each team masters its domain, uses advanced tools, and operates with a high level of sophistication. But each plays its own score, often on its own tempo.

What is missing is not talent or information: it is orchestration. The role of the conductor is not to replace musicians, but to align them. To ensure that timing, intensity, and interactions produce a coherent whole. Without this role, individual excellence does not translate into collective performance.

In companies, this orchestration is partially assumed through leadership, governance, and processes. But as complexity increases, this approach reaches its limits. The volume and interdependence of decisions make coherence impossible to manage manually. This is why organisations need a more systematic form of orchestration – one that combines data, technology, and human judgment into a coherent system. This is not a new idea, but it is becoming unavoidable at scale.

‍

A new source of sustainable competitive advantage

What is emerging is a different way of thinking about performance. Companies rarely fail because of a single bad decision, but because of a sequence of misaligned decisions that accumulate over time. Conversely, sustained performance is not the result of isolated successes, but of decisions that consistently reinforce each other.

For years, companies have invested in becoming insight-driven. The next step is to become decision-driven, and ultimately outcome-led. The competitive advantage shifts accordingly. It no longer comes primarily from having better data or better models. It comes from the ability to organize decisions so that they reinforce each other and consistently translate into business impact. This means leveraging data and AI to unlock the right insights, transforming these insights into individual decisions, then orchestrating these individual decisions within a winning decision system for the company.

This requires more than technology alone. It requires a combination of systems and capabilities. A platform to structure decision-making, ensure scalability, and connect insights to execution. And the expertise to understand business context, design relevant use cases, and navigate the complexity of real-world implementation.

‍

Toward a decision system for 21st century organizations

What does a decision system for 21st century organisations look like? In practice, it means creating a shared language of performance across functions, aligning KPIs and trade-offs between finance, marketing, and commercial teams, and embedding decision logic directly into workflows. It means connecting data, models, and business rules into a coherent system that can simulate, arbitrate, and execute decisions at scale.

This has been the logic behind our work for the past 20 years. Not as an additional layer of analytics, but as an operating system for decision-making, built from two decades of working at the core of large organizations, observing how decisions are made, where they fail, and how they can be structured to deliver consistent outcomes. It integrates platform and expertise to ensure that insights translate into aligned decisions, and decisions into measurable impact.

In that sense, performance is no longer driven by isolated decisions, but by the ability to orchestrate them – consistently, at scale, and in service of outcomes.

‍

To read more about this subject, explore more articles here and here.