Priya is a senior claims handler at an insurer and works from home three days a week. In the spring the company installed software that produces a daily “activity score” from keyboard input, mouse movement and the applications open on her screen. Her score has slipped into the bottom tenth of her team. The dashboard shows long idle periods most afternoons.
Those afternoons are when she takes the complicated cases: reading paper medical reports and phoning claimants on the desk handset. None of it registers. Her manager, Tom, under pressure to “act on the data”, invites her to an investigation meeting about “time recording”. Within a week two colleagues have bought mouse jigglers and nobody is volunteering for the difficult cases any more.
Employee monitoring software, often called bossware, has spread steadily since the pandemic. It is now a discipline issue as much as a privacy one: dashboards produce numbers, numbers look like evidence, and evidence ends up in disciplinary files.
This guide looks at how surveillance changes behaviour, why discipline built on activity data so often backfires, and what managers can do differently. It sits alongside our guides to why punishment rarely changes behaviour and organisational justice.
Monitoring measures activity, not work.
The more a number is used to judge people, the more people work on the number.
Surveillance shifts responsibility.
People who feel watched tend to hand their conscience to the watcher and behave worse, not better.
Purpose, limits and voice decide the outcome.
Monitoring that is explained, narrow and agreed is tolerated. Monitoring that feels like a trap breeds conflict.
Why is monitoring in the news now?
Because regulators in several countries have acted on it at once.
In California, Governor Newsom signed four workplace technology laws on 30 September 2026. According to a summary by the law firm Epstein Becker Green, SB 947, the No Robo Bosses Act of 2026, will from July 2027 stop employers relying solely on an automated system to discipline or dismiss someone; where its output is relied on, a human must corroborate it. AB 1883 bans AI surveillance tools that infer employees’ emotional states or collect neural data, and AB 1331 bans surveillance in restrooms, both from January 2027.
Connecticut’s expanded electronic monitoring law took effect on 1 October 2026, requiring employers to say where on their premises monitoring may happen and to give new hires a plain-language statement of what may be monitored. In the UK, the Department for Business and Trade ran a consultation on workplace monitoring technologies from July to September 2026, with options ranging from guidance to a legal duty to consult workers before introducing certain tools.
The American Psychological Association’s Monitor on Psychology ran a November 2026 feature, “The hidden costs of workplace surveillance”, arguing that expanded monitoring can undermine performance and weaken trust.
What does the evidence say monitoring does?
Less for performance than its buyers hope, and more to wellbeing than they expect.
A meta-analysis of 70 samples found that electronic monitoring was associated with slightly lower job satisfaction and slightly higher stress, had essentially no relationship with performance, and had a small positive relationship with counterproductive work behaviour. Performance targets and feedback attached to the monitoring appeared to make the negative effects worse.
The effects are small, but the direction is consistent, and the missing performance benefit matters, because productivity is the usual justification.
The APA’s 2023 Work in America survey, conducted by The Harris Poll with 2,515 US employees, found that just over half said their employer used technology to monitor them. Of those monitored, 56% said they typically felt tense or stressed at work, against 40% of those not monitored, and 51% felt micromanaged, against 33%.
The US Government Accountability Office reviewed 122 studies published between 2020 and 2024. It found that digital surveillance can protect workers, for example by flagging safety risks, but can also raise injury risk when it pushes people to work faster to meet metrics. Several studies linked a lack of transparency about how data is used to worker anxiety.
Why does being watched make behaviour worse?
Three mechanisms explain most of it.
Responsibility moves to the watcher. In research reported in Harvard Business Review in 2022, David Welsh of Arizona State University and colleagues found that monitored employees were more likely than unmonitored ones to take unapproved breaks, ignore instructions and work slowly on purpose. Their explanation: monitoring weakens people’s sense of agency. When someone else is checking, being good feels like the watcher’s job.
Autonomy is a basic need. Self-determination theory holds that people need to feel some control over how they work. Close monitoring signals that they are not trusted with that control. The usual reaction, reactance, is to reassert freedom wherever possible, sometimes by quietly defeating the system.
People work on the measure. When an activity score is used to judge performance, the rational response is to raise the score. Mouse jigglers and scheduled emails are not unusual dishonesty; they are what anyone does when only the measure is rewarded.
A rising activity score after monitoring is introduced looks like proof the tool works. Often it shows only that people have learned what the tool counts, while real output stays flat.
What happens when oversight runs on fear?
An unusual natural experiment is under way in China. Writing in Human Resources Director in October 2026, Matthew Sellers described research by Ben Hillman and Tan Zhao of the Australian National University, published in The China Journal in January. Drawing on fieldwork in six provinces, they found that local officials facing stricter discipline, heavier reporting and elaborate performance reviews often responded with “formalism”: shirking, feigned compliance, falsification and form over substance.
The parallel with bossware is direct. As the article concludes, oversight that people trust tends to make an organisation safer, while oversight they fear mostly teaches them to look busy.
Why does dashboard-driven discipline produce conflict?
Because it breaks almost every rule of fair process at once.
The data lacks context. A log of idle minutes records what a keyboard did, not what a person was doing. Discipline based on a measure the employee knows to be wrong is experienced as unjust from the first letter.
The employee had no voice. Research on procedural justice consistently finds that people accept unwelcome decisions more readily when they had a say in the process. Monitoring that arrives without consultation, then generates allegations, denies voice twice over.
Automation bias. People tend to over-trust figures produced by a system, especially under time pressure. An October 2026 article in CIO, reporting on the California law, made the point in its headline: a human rubber stamp on AI-based decisions will not satisfy it. A manager who simply signs off what the system flagged is not exercising judgement.
Everyone is watching. A disciplinary case built on monitoring data is a public lesson in what the organisation values. Priya’s colleagues concluded that visible activity counts and slow, careful work is risky. One case becomes a team-wide conflict with management.
Monitoring data introduced for one purpose and later used for another, such as security logs reused to build a conduct case, is among the fastest routes to a grievance. It confirms the suspicion that the tool was always a trap.
What managers can do
Monitoring is sometimes justified for safety, security or regulatory reasons. The question is how to use it well.
Treat data as a question, not a finding. If a dashboard suggests a problem, start with an open conversation, not an investigation. “Your activity score is low in the afternoons. Talk me through what those hours usually involve.” Priya would have had a good answer. The guide to difficult conversations covers how to open such a discussion without triggering defensiveness.
Judge output, not motion. Check quality, outcomes and deadlines first. If they are fine, the activity data is telling you about the tool, not the person.
Never let a metric be the only evidence. Corroborate with other sources and give the employee the full data and a chance to explain it. Fair process requires it, and regulation is heading the same way.
Separate monitoring from performance plans. Putting someone on a performance improvement plan with a target activity score invites gaming and tells them the plan is about appearances.
- State the specific problem it solves, and check employees would describe its purpose the same way
- Collect the least data that solves that problem
- Write down what the data will never be used for
- Consult employees or their representatives before deciding
- Require human review, with context, before any data feeds a disciplinary step
- Review in a year whether it is still used only for its stated purpose
Ask your team, anonymously, one question: “Does the way we measure work ever push you to do something that looks productive rather than something that is?” If many say yes, the measure is producing theatre, and no amount of discipline will fix that.
How did it end for Priya?
The meeting did not go ahead. HR asked Tom what evidence he had beyond the dashboard, and he had none; her closure rate on complex claims was the highest on the team. Tom apologised, asked how her afternoons worked, and the team agreed to log complex cases by outcome and keep activity scores out of one-to-ones. The software had been sold as a way to build trust in remote working. Used as a disciplinary tool, it did the reverse within a month.
Frequently asked questions
What is bossware?
An informal name for software that tracks employees’ digital activity, such as keystrokes, applications, screenshots or location, often summarised as an activity score.
Does employee monitoring improve productivity?
The research gives little support for that. A 2022 meta-analysis found no relationship between electronic monitoring and performance, alongside small increases in stress and counterproductive behaviour and a small fall in job satisfaction.
Can monitoring data be used in a disciplinary case?
Often it can, subject to local law, notice requirements and data-protection rules, which are tightening in places such as California and Connecticut. Fair process requires that it is corroborated, shared with the employee and interpreted in context rather than treated as conclusive.
How can monitoring be introduced without damaging trust?
State its purpose, collect only what that purpose needs, consult employees first and commit in writing to what it will not be used for.
Discover more from Kurums | Business Intelligence
Subscribe to get the latest posts sent to your email.