Care in an organisation does not mean knowing everything about every person or turning each minute, message, and pause into evidence of performance. A system can prevent overload, make support available, and make work safer without continuously recording presence, speed, connections, or intimate circumstances. The deciding criterion is purpose: collect only information needed for a specific protection, use it only for the necessary time, and prevent a helpful measure from becoming concealed disciplinary control. This text sets out general organisational and protective criteria; it is not legal, employment, medical, or human-resources advice for individual circumstances.
Care does not require an all-seeing view#
The promise to monitor everything is often presented as efficiency. A dashboard combines activity, availability, deadlines, location, or keystrokes and claims that it will detect exhaustion earlier. Yet abundant signals are not necessarily useful signals. They can confuse presence with contribution, penalise rhythms compatible with care, illness, or disability, and replace a conversation about resources with individual comparison. A responsible organisation first asks which risk it needs to prevent and what minimum evidence permits action. If the problem is impossible planning, counting clicks does not solve it; objectives, staffing, priorities, and recovery time need review.
The promise to monitor everything is often presented as efficiency. A dashboard combines activity, availability, deadlines, location, or keystrokes and claims that it will detect exhaustion earlier.
Privacy is also a practical condition of safety. Someone who fears that a question, a pause, or flexible hours will be interpreted against them is likely to hide difficulty. That silence degrades the information an organisation needs in order to care. Trust is not the absence of rules; it is the predictability that rules will not be used beyond their purpose. The data collected, the reason, access, decisions influenced, and route for challenging an error should be explained. Information about health, family life, location, or communications needs heightened care and should not circulate for managerial convenience.
Purpose, necessity, and proportionality#
Before enabling a tool, it is useful to complete a short record: legitimate objective, risk to be reduced, less intrusive alternatives, strictly necessary data, responsible roles, retention period, and planned review. This discipline stops a temporary function becoming permanent through inertia. Proportionality requires comparing expected benefit with intrusion and secondary effects. A solution that helps address a defined incident may be reasonable; one that records an entire workforce without pause in order to anticipate unlikely cases usually is not. Assessment should include people doing the work and, where relevant, worker representation and prevention functions.
Aggregated data often support better decisions than individual tracking. To review workloads, an organisation can examine work volume, shift changes, recurring incidents, and coordination time without building a profile of every person. To assess training, it can use process outcomes and spaces for questions instead of capturing screens or measuring every connection. Where identifying a case is truly necessary, access should be limited, logged, and supported by a verifiable reason. Minimisation is not paperwork: it reduces leaks, interpretive errors, and the power to reuse old data for new purposes.
Availability is not a measure of commitment#
Digital channels have made visible activity that once left no trace: last connection, response speed, application status, or a shared calendar. Turning those traces into a constant expectation changes the working day even when nobody gives an express instruction. A person answers during a meal, delays an appointment, or avoids a break because silence may look like a lack of commitment. Organisational care protects disconnection through understandable hours, explicit on-call rotas, cover arrangements, and criteria for emergencies. It also prevents a private phone from becoming a compulsory doorway to work.
Case 1: the remote team and the green indicator. A unit observes that people shown as less available in chat close fewer tasks. The first response is to demand visible connection throughout the day. Reviewing the process reveals that several tasks require concentration, calls with users, and work in a system that does not change the indicator. Rather than watch status, the team agrees response periods, an emergency channel, and objectives reviewed for outcome and quality. The measure reduces interruption and identifies genuine delays without making a screen a test of loyalty.
Workloads are corrected with resources, not profiles#
Excessive workload is rarely explained by one behaviour. It can arise from an uncovered absence, an unrealistic timetable, a slow tool, coordination that is not counted as work, or an unequal allocation of invisible tasks. Measuring who takes longer can reinforce the problem when those conditions are not examined. Management needs the capacity to postpone delivery, simplify steps, add support, and redistribute responsibility. No metric replaces the decision to reduce a demand that is incompatible with a safe working day.
Case 2: the connected-hours alert. A service introduces automatic alerts when someone exceeds a number of activity hours. The alert goes to their manager with a daily connection detail. Several people stop recording tasks so as not to appear slow, while others remain connected because they fear explaining an impossible workload. Review changes the mechanism: aggregate process trends are examined, the team can declare obstacles without intimate detail, and a regular meeting decides priorities. Individual alerts are reserved for a voluntary request for support or a clearly defined safety situation.
Design systems that are secure and challengeable#
A secure system does more than protect passwords. It restricts access, separates duties, retains access records, and allows incorrect information to be corrected. If a tool recommends task allocation, shift selection, or assessment, its relevant criteria, limits, and the person authorised to review its result must be known. Automation may organise information, but it should not make unavoidable a decision affecting income, time, or reputation without a meaningful opportunity for explanation. Exceptions must be possible and must not become a cost borne only by people with less power.
Case 3: the shift platform. An algorithm assigns preferences from historical availability. People who reduced hours for care appear to have fewer options and repeatedly receive worse shifts. The supplier offers a technical report but no usable route for correction. The organisation stops the rule, reviews variables, and creates an appeal handled by people with a response deadline. It also limits data to what is needed for the current shift and tests the system with varied situations before restarting it. Care here means preventing a trace of the past from determining the future indefinitely.
Governance, participation, and learning#
Decisions about monitoring cannot rest only with the person buying technology. They need an identifiable owner, an inventory of tools, periodic impact review, and a protected route for raising concerns. Affected people know tasks, workarounds, and workloads that do not appear in a sales presentation. Their participation improves design and reduces the risk that an apparently neutral system penalises those working at different rhythms. When a measure is changed or removed, it is helpful to communicate what was learned without disclosing personal data.
Training should explain limits as well as duties. A platform administrator needs to know that curiosity does not justify access; a team coordinator needs to distinguish a supportive conversation from collection of sensitive information; and a worker needs to know which route can correct a record or report misuse. Incidents require a response: contain access, inform people where appropriate, repair harm, and redesign the control. Concealing a failure to protect image destroys the trust that care was intended to create.
A practical test before and after#
Before adopting any monitoring, ask: what specific decision will it enable; could that decision be made with less data; what happens to people who cannot always be visible; who will review a wrong conclusion; and when will records be deleted? Afterwards, check whether risk fell without fear rising, whether resources actually changed, and whether there is evidence of exclusion or retaliation. A change that only makes the workforce more legible, rather than work more sustainable, is not care.
Applicable rules vary by territory, sector, and employment relationship, so particular cases require competent advice. As general references, the International Labour Organization addresses occupational safety and health; EU-OSHA gathers resources on psychosocial risks and digitalisation; Spain's National Institute for Safety and Health at Work publishes prevention materials; and the Spanish Data Protection Agency addresses data processing in employment relations. The organisational question remains: does the information collected permit concrete, limited protection, or does it merely increase the capacity to watch?
Review must also attend to cumulative effects. A single question about availability can look harmless, but a sequence of requests, reminders, and classifications creates an expectation that shapes everyday choices. It is necessary to review not only each data field but the whole set of practices: meetings, out-of-hours messages, reports, captures, and automatic rules. When several sources can reconstruct a working life in detail, the risk exists even if none is described as monitoring. Limiting that combination, setting short retention periods, and independently checking actual use are verifiable forms of care.
Meaningful accountability includes a decision log that people can understand. It should say why a monitoring measure was selected, what alternative was rejected, which team can access the information, and the date on which necessity will be reconsidered. Regular review should seek disconfirming evidence, including reports from people who experience a tool differently because of role, disability, caring responsibilities, language, or location. A system is not justified merely because it is familiar or because a supplier calls it standard. Its legitimacy depends on a continuing demonstration that it protects a real need with the least intrusive workable means.
Such boundaries make support credible, understandable, and safer for everyone affected by work.
Institutional sources: International Labour Organization, occupational safety and health; EU-OSHA, psychosocial risks; INSST, psychosocial risks; Spanish Data Protection Agency, employment relations and personal data; European Commission, data protection.
