Sensing the Conditions

When the questions fall silent

On Friday afternoon, a distribution company changes a pricing rule. Over the weekend, an AI agent processes thousands of customer requests using the old rule. Nobody updated its instructions. By Monday, the work is done, consistently and wrongly.

A person might have noticed the mismatch, asked in the team chat or called the pricing lead. Slightly inefficient, perhaps irritating, and extremely valuable. The question was the signal.

People have always revealed something about organisational conditions through their work. They ask when priorities are unclear, struggle when capability is missing, and push back or go quiet when the climate deteriorates. Healthy organisations use these signals to learn. As AI agents take on more work, those signals may grow quieter. Even healthy organisations will need to rebuild them deliberately.

From periodic readings to continuous attention

Throughout this series, conditions means three things: Clarity, the direction, priorities and responsibility for decisions; Capability, the competence and capacity to deliver; and Climate, whether people feel safe and willing to contribute their judgement.

Execution is becoming continuous, while organisations still tend to read these conditions periodically. Strategy reviews happen quarterly, engagement surveys annually. Smaller companies often rely on proximity rather than formal instruments, leaving them particularly exposed when everyday human signals disappear.

The interval between readings was never a principle. It reflected the cost of collecting information, interviewing people and interpreting results. AI reduces that cost. Exception logs, project histories and customer conversations can now be analysed more often and together. Patterns can emerge that nobody thought to put on a dashboard.

I call an organisation that deliberately uses this capacity The Continuum Organisation™. Its analysis of conditions is no longer an event but a continuum. This means continuity of attention, not permanent organisational change.

Its purpose is to become healthier: notice what works and spread it, offer support while helping is still easy, and adjust direction through small steps. A barometer, not an alarm.

Headlights as well as brakes

Human governance provides the brakes that make speed possible. Sensing adds the headlights. At night, safe speed depends on how far a driver can see.

Organisations often hesitate to expand AI because leaders cannot see whether their conditions will hold. Continuous sensing allows them to move faster where conditions are strong and invest where support is needed, team by team.

The principle remains: stability precedes speed. But stability becomes something maintained while accelerating, rather than a phase completed beforehand.

Four ways to read conditions

Read strength, not only direction. If 78 per cent of employees say they understand the strategy, that might reflect conviction or simply the expected answer. Look for consistency between stated agreement and behaviour. Where greater depth is useful, voluntary response-time methods, reported only for groups, can add another reading of how firmly a position is held.

Read the spread, not the average. Conditions are experienced locally. An organisation-wide average can hide both a struggling team and an exceptionally healthy one. Differences reveal where support is needed and where there is something worth learning from. A team doing unusually well shows what good conditions look like in your own setting.

Read what everyday work already shows, before asking. Repeated escalations, the movement of ideas between teams and delays in passing on bad news all carry information about conditions. These signals arise through work rather than through answers to a questionnaire. Reading them still requires transparent boundaries.

Let people interpret their own experience. Short accounts of everyday events can reveal what formal measures miss. People should make sense of these accounts themselves, rather than surrendering interpretation to an expert or a machine. AI can find patterns across many accounts; meaning remains with the people who contributed them.

Climate is especially difficult to read because the instrument is affected by what it measures. People who feel unsafe speaking up may not say so in a survey. High scores can conceal quiet compliance. Everyday signals and people's own interpretations therefore matter particularly here.

AI widens attention; people give it meaning

There is an apparent contradiction: if AI can miss its own limits, why trust it to sense organisational conditions?

The answer is a division of labour. AI can examine more material and surface patterns. People must recognise when the interpretation no longer fits. Some problems yield to analysis; questions of trust and culture often require small, safe-to-fail interventions and careful attention to what follows.

Machine signals are also incomplete. An agent may not recognise that its instructions are inadequate. The evidence may appear elsewhere, in customer complaints, overrides or rework. Human and operational signals must be read alongside machine ones, without assuming automation knows best.

The sensing system itself needs review. People must periodically reconsider whether its signals still mean what they think they mean and whether it is looking in the right places.

More analysis does not mean more judgement

Attention remains finite. Choosing what to read is a strategic decision about what matters and what the organisation can change. Analysing everything is not the same as understanding it.

A useful sensing system learns. It drops signals that never affect a decision and adds those whose absence became apparent. Recording the context and interpretation behind decisions makes later learning possible without turning hindsight into blame.

AI can help turn findings into something usable. Strategic decisions, however, bind people who did not make them. They must remain with accountable humans.

Sense with people, not about them

The boundary is simple: individual insight belongs to the individual; organisational insight diagnoses conditions, not people.

People may receive personal results from information they knowingly contribute. The organisation should see patterns across teams and over time, not named profiles for ranking or judging individuals. No hidden profiling.

Teams need to know what is being read and why, and see their own readings first. Surveillance would damage the very climate sensing is meant to protect.

The Continuum Organisation makes learning from conditions deliberate rather than accidental. It does not settle who may act, under which mandate or with which stopping rules. Those questions lead towards AUREA™, the Strategic AI Operating Blueprint. Sensing provides the headlights. Responsibility for the wheel remains to be designed.