On living systems

Abstract

When outcomes can’t be predicted in advance, organisations must rely on sensing, feedback and adaptation. These are properties of living systems, not machines.

A post-predictive world

For much of the industrial era, organisational success rested on the ability to predict outcomes with reasonable confidence. Markets moved slowly enough to be forecast. Demand could be modelled. Cause and effect were sufficiently separated in time to allow for planning, execution, and correction.

Under these conditions, prediction was not an aspiration. It was a practical capability.

That capability is now increasingly fragile. Organisations operate in environments shaped by interacting forces whose effects are difficult to isolate or anticipate. Change unfolds unevenly. Small signals can cascade into large consequences. Outcomes often become clear only after action has already been taken.

This does not mean organisations face more uncertainty than before in absolute terms. It means that uncertainty has become structural. The future is no longer something that can be reliably inferred from the past.

When outcomes cannot be predicted in advance, the role of planning itself must be reconsidered.

The machine logic organisations still rely on

Despite this shift, many organisations continue to operate as if prediction remains possible.

Their structures reflect a machine logic: objectives are defined upfront, plans are created to achieve them, execution is monitored against targets and performance is optimised through control and efficiency. Inputs are transformed into outputs through processes designed to minimise variance.

This logic is deeply embedded. It shapes governance, budgeting, performance management and the language organisations use to describe progress. It assumes that the organisation can specify what it is trying to achieve and then align effort accordingly.

This was not a naïve assumption. It was a historically successful one. Machine logic delivered scale, consistency and reliability in environments where outcomes could be defined with confidence.

The difficulty is that this logic presumes conditions that are increasingly absent.

Why machine logic fails under unknowable conditions

Machines function well when outcomes are known or can be specified in advance. They are designed to execute, optimise and repeat. Their strength lies in precision and control.

Under unknowable conditions, these same strengths become constraints.

Such optimisation locks behaviour around assumptions that may no longer hold. Control slows response as authority concentrates. Execution accelerates action before understanding has stabilised. Efficiency reduces redundancy that might otherwise absorb shocks or surface weak signals.

The organisation becomes highly capable at doing the wrong things well.

This is not a failure of discipline or competence. It is a mismatch between system design and environmental reality. Machine logic depends on predictability. When predictability disappears, the organisation expends increasing effort maintaining control while becoming less responsive to what is actually happening.

The issue is not that planning is obsolete. It is that planning can no longer be the primary mechanism for navigating an unknowable future.

Systems designed to remain viable without prediction

At this point, a different class of systems becomes relevant.

Living systems persist without knowing outcomes in advance. They do not operate by specifying end states and optimising toward them. Instead, they remain viable by continuously sensing their environment, responding to feedback and adapting their behaviour over time.

This is not a metaphor. It is a description of how systems function when prediction is unreliable.

Living systems maintain coherence not through control, but through responsiveness.

They adjust based on what they encounter rather than what they expect. They do not eliminate uncertainty; they operate within it.

Crucially, they do not require a stable future to function. They require only the capacity to sense, respond and adapt as conditions change.

When organisations face environments that cannot be reliably predicted, these properties become relevant not as inspiration, but as necessity.

Sensing, feedback and adaptation as system properties

In systems designed for unknowable conditions, sensing takes precedence over forecasting. What matters is not the ability to predict outcomes, but the ability to notice change early and interpret its significance.

Feedback becomes more important than targets. Rather than measuring performance solely against predefined objectives, the system pays attention to the effects of its actions and adjusts accordingly. Learning is continuous rather than episodic.

Adaptation replaces optimisation as the organising principle. Instead of refining processes toward an assumed end state, the system remains flexible, adjusting behaviour as understanding evolves.

These properties are not tactics. They are characteristics of systems whose primary concern is ongoing viability rather than short-term efficiency.

In organisational terms, this shifts the focus from executing plans to maintaining alignment with a changing environment.

The persistence of machine responses

Despite recognising these dynamics, organisations often respond to uncertainty using familiar tools.

Agility initiatives promise faster execution. Resilience programmes focus on recovery. Transformation efforts define future states and roadmaps to reach them. Each of these responses assumes that the organisation can specify where it is going and then reorganise itself accordingly.

These approaches are understandable. They offer reassurance and structure in the face of ambiguity.

The difficulty is that they remain rooted in machine logic. They accelerate movement without increasing understanding. They optimise change rather than learning. They assume that adaptiveness can be installed rather than cultivated.

As a result, organisations often move quickly while remaining misaligned with their environment. Change becomes performative. Effort increases without corresponding gains in responsiveness.

The problem is not a lack of initiative. It is the continued reliance on models that presume predictability.

Reframing the organisational challenge

If outcomes cannot be known in advance, and if machine logic constrains responsiveness under such conditions, then the organisational challenge must be reframed.

The question is no longer how to execute better plans, but how to remain aligned with a reality that cannot be fully anticipated. This requires systems that prioritise sensing, feedback, and adaptation over control and optimisation.

Such systems behave less like machines executing instructions and more like living systems maintaining viability. They do not abandon discipline or purpose, but they organise around responsiveness rather than prediction.

This reframing does not provide answers. It exposes a tension.

Many organisational assumptions about performance, control and change were formed in environments where outcomes could be specified in advance. As that assumption weakens, so too does the logic that flows from it.

What it would mean to design organisations that function effectively under these conditions remains unresolved. What is increasingly clear is that when outcomes cannot be predicted, organisations must rely on properties that machines do not possess.