An organisation's costliest decision is often made before a tender opens, a meeting is called, or an application is switched on. It happens when somebody names a problem without listening to the people who know it from inside. A tool can speed calculations and arrange tasks, but it cannot decide which loss matters, what risk is acceptable, or who will carry additional work. Responsible leadership deliberately delays the answer in order to bring the question forward. It asks what is happening, who is affected differently, what has been tried, and which relationship deserves protection. That move does not weaken authority. It makes authority intelligible, because a private intuition becomes a decision others can examine, amend, and support.
Define the problem before requesting a solution#
Meetings frequently begin with a practical-sounding statement: we need a platform, protocol, dashboard, or campaign. The statement already contains a wager and excludes alternatives that may address the situation better. Careful leaders replace it with observable questions. Where does the process stall? Who waits longest? What information is missing? What result would demonstrate improvement? The change is modest but it separates a symptom from its cause. If a library has fewer visitors, the issue may be its opening hours, signage, programme, or trust; purchasing an app does not automatically answer any of those questions.
Meetings frequently begin with a practical-sounding statement: we need a platform, protocol, dashboard, or campaign.
Good definition needs several kinds of evidence. Usage data reveal patterns, conversations reveal reasons, and observation exposes friction nobody records. It also requires stating what remains unknown. A percentage does not explain why someone abandons a form or whether another person was unable to begin it. The team should write a provisional hypothesis, identify who will challenge it, and set a review date. Precision is not pretending certainty; it is leaving the route from experience to interpretation and action visible. When that route is shared, an organisation is less likely to mistake the preferences of its loudest members for the needs of an entire community.
Listening is design work, not a ceremony#
Listening is not collecting opinions to validate a decision already made. It must generate information that can alter scope, timing, resources, or the wisdom of the project itself. That means inviting service users, maintainers, and people most readily excluded. Invitations should say what is still open, which boundary cannot change, and how replies will be handled. A useful survey accepts stories as well as closed choices. A useful conversation offers accessible times, interpretation where needed, and a note-taker who does not assign blame. Inclusive participation takes time, sometimes compensation, and channels that do not rely on one device or one confident style of speaking.
Silence is information too, but it must not be read conveniently. It may indicate agreement; more often it reflects fatigue, previous consultations that changed nothing, language barriers, or fear of contradicting the person in charge. Leaders should actively seek those absences rather than asking one individual to stand for a whole group. Short interviews, open sessions, workplace visits, and incident reviews complement one another. The team then returns what it heard in plain language: areas of agreement, continuing disagreement, and proposals it cannot adopt. This return closes the listening loop and lets participants test whether their experience was understood before it is turned into a requirement.
Case one: the shift the algorithm could not see#
A care cooperative planned to introduce automatic shift allocation. Its dashboard showed unexpected absences, and managers assumed the problem was discipline. Before buying the system, a coordinator spoke with care workers, families, and administrators. She found that changing addresses, late notice of hospital discharges, and public-transport journeys made it impossible to follow a route calculated as though every street and every day were alike. Workers were also exchanging support by phone, cooperation invisible in the records. The first question changed from “how do we reduce absences?” to “how do we design shifts without moving uncertainty onto the people who provide care?”
The response combined a reserve period, one channel for urgent changes, and rotating worker participation in route review. It was tested for six weeks in one area with an explicit option to stop. The cooperative tracked delays, overtime, continuity for families, and the emotional burden reported by staff. The automatic system remained a way to suggest alternatives, not a judge of availability. Last-minute changes fell, but the crucial lesson was that decisive knowledge lived in daily work, not the spreadsheet. Management gained a rule for future purchases: no model is adopted until people who will absorb its errors have tested its assumptions.
Turning disagreement into a legitimate decision#
Collective decision-making does not require unanimity. It requires that disagreement has an understandable place and that final criteria are public. When two options compete, compare them through the same frame: purpose, affected people, total cost, accessibility, environmental impact, risks, reversibility, and owner of each consequence. The frame stops the proposal with the most slides from appearing inevitable. It also requires serious consideration of waiting, simplifying a process, or strengthening existing practice. Taking time over definition usually saves time in repair. Speed is a virtue only when it does not purchase silence, exclusion, or a debt that somebody else must pay later.
The accountable person must ultimately say what will be decided and why. Delegating listening does not mean dissolving responsibility into a confusing vote. A decision note can fit on one page: the problem, evidence consulted, options considered, criterion that tipped the balance, foreseeable effects, review date, and route for challenge. This record protects both the community and institutional staff. It helps reveal when an exception becomes a rule and allows results to be discussed without relying on meeting memories. Useful transparency is not publishing every document; it is showing the reasoning that makes a choice challengeable without requiring people to guess how it was reached.
Case two: a school and families' time#
A school planned to replace face-to-face appointments with digital booking to improve communication. Its leadership assumed families would welcome convenience. Conversations with the family association disclosed a different reality: some people shared a phone, others worked changing shifts, and several needed to discuss sensitive matters without leaving a trace on a platform. The real problem was not the absence of an electronic diary. Existing routes forced many families either to seek permission at work or to explain their circumstances in front of other people. The school stopped presenting the platform as a universal answer and made accessibility a condition of success.
It retained phone calls, arrival and departure appointments, and a weekly slot outside normal hours; online booking became one additional door. Two mediators reviewed message language with staff and recorded, without unnecessary data, which channel enabled each contact. Attendance at tutorials rose after two months, not because of technology itself but because each family could choose a dignified way to participate. The case shows that the same tool can widen or narrow a relationship depending on its surrounding framework. Leadership meant accepting that efficiency is not moving everybody through the institution's cheapest channel.
Test, care, and correct#
A new decision should begin with limited scope and a safe exit. A pilot is not permission to experiment without protection: it needs a purpose, suitable consent, agreed indicators, and a stopping procedure. Alongside performance metrics, reviewers should observe how burdens are distributed, errors, complaints, accessibility, and trust. An average improvement can conceal a group's loss of time, privacy, or ability to understand what is happening. Every review should therefore ask who receives the benefit, who bears the cost, and who can request repair. Without a specific answer, conditions do not yet exist for expansion.
Correction is not a moral confession of failure. It is the obligation of a person who retains decision-making power. When harm appears, the priority is protecting the affected person, providing an equivalent alternative, and explaining the change. Afterwards, investigators should ask whether an ambiguous instruction, excluding interface, badly designed incentive, or inadequate information made the problem predictable. Blaming the last person in the chain prevents learning. Mature organisations retain a record of incidents and adjustments to train new teams and detect patterns. Accountability becomes credible when changing course is possible, funded, and visible to the people who raised the concern.
Case three: the park that needed conversations#
A city council received a proposal to place sensors and cameras in a park after safety complaints. The offer promised real-time alerts and maps of problematic zones. Before approving it, the community team walked the site with young people, shopkeepers, older residents, cleaners, and disability groups. Their conversations showed that poor lighting, closed toilets, and a lack of evening activity explained much of the unease. Some participants also feared indiscriminate surveillance would make their presence appear suspicious. The question became not “how do we monitor better?” but “what conditions allow different people to use this park with confidence?”
Investment was divided between repairable lighting, open facilities, community programming, and mediation training. Any future sensor would require a citizen panel, privacy assessment, and public evidence of usefulness. After a season, reports fell and park use increased, although no single cause can claim every change. The important result was that residents could see what had been decided, take part in review, and question an intrusive option before it became difficult to remove. Listening did not create a spectacular answer. It created one proportionate to the place, open to revision, and closer to the life already taking place there.
Make the first question a habit#
Making the first question routine means preparing the organisation before urgency arrives. A short prompt for every initiative can describe the problem without naming a solution, identify unequal effects, invite relevant voices, compare alternatives, and publish a review commitment. Middle managers need permission to raise doubts and protected time to speak with people delivering the service. Users need to know where to propose, how to appeal, and what response time to expect. No form alone can replace this architecture of trust. The appropriate tool arrives afterwards, once its function and limits can be explained without vague promises.
Collective leadership does not refuse to decide; it refuses the illusion that deciding alone is more objective. Its strength lies in giving dispersed knowledge consequences and making power answer for outcomes. Before purchasing, automating, or announcing, return to the opening question: what reality are we trying to improve, and who must be able to dispute our answer? If an organisation can respond with evidence, care, and a real correction route, it is ready to select a tool. If it cannot, continuing to listen may be its most responsible action. Decision quality is recognised not by launch speed but by the dignity shown to people who make it possible.
Institutional sources for further work#
The OECD collects work on open government; UNESCO provides its recommendation on the ethics of artificial intelligence; the World Health Organization publishes resources on safety and participation; and the European Commission explains public consultation mechanisms. They are institutional starting points, not substitutes for the situated judgement of each community.
