Algorithmic management in European workplaces

Algorithmic management is the use of digital technologies to fully or partly automate managerial functions, including the organisation, coordination, monitoring, supervision and evaluation of work. While originally associated with digital labour platforms, the term is now understood as applying more broadly across a wider range of employment settings.

Although algorithmic management can enhance efficiency and risk management and, in some sectors, such as financial services and transport, may be driven by regulatory compliance, it also presents challenges and potential drawbacks. While quantitative evidence based on representative surveys is comparatively limited, there is a growing body of research – largely based on qualitative studies – that links algorithmic management to work intensification, increased time pressure and reduced worker autonomy. Existing evidence also suggests that the effects are contingent on the job resources available to workers, such as autonomy, organisational support and skills use, which can buffer potential adverse impacts.

This report contributes to the ongoing policy debate by presenting findings from the analysis of European Working Conditions Survey 2024 data, with a focus on algorithmic management in European workplaces. In this report, algorithmic management is measured by whether employees report that computer programs are used at work to allocate their tasks, determine their working hours (scheduling) or monitor their performance. This simplified operationalisation captures three core management functions of algorithmic management but does not provide an exhaustive definition of the concept. The analysis is restricted to employees only, examining both the prevalence of algorithmic management and its relationship with working conditions and employees’ health and well-being.

To investigate the combined association of algorithmic management, job resources and job demands on health and well-being outcomes, this report draws on the concept of job strain, defined as a situation in which job demands exceed job resources. Employees are distributed across a range of job-quality profiles from highly strained jobs characterised by high demands and limited resources to highly resourced jobs characterised by a more favourable balance between demands and resources.

Following the adoption of the Platform Work Directive (Directive (EU) 2024/2831), policy attention has progressively moved beyond its original focus on algorithmic management in platform work towards its use in more conventional work settings. The classification of certain artificial intelligence (AI) systems used for work management as high-risk systems under the Artificial Intelligence Act (Regulation (EU) 2024/1689) marks an important step in the EU’s approach to regulating digital technologies in the workplace and reflects growing concern about the implications of these technologies for employees.

Other EU legislative instruments that are relevant to algorithmic management include the General Data Protection Regulation (Regulation (EU) 2016/679), which establishes safeguards for the processing of workers’ personal data and automated decision-making; the Information and Consultation Framework Directive (Directive 2002/14/EC), which provides for the involvement of workers’ representatives in decisions affecting work organisation; and the Occupational Safety and Health Framework Directive (Council Directive 89/391/EEC), which requires employers to assess and prevent risks to workers’ health and safety, including psychosocial risks.

More recently, the European Parliament, in its resolution on digitalisation, artificial intelligence and algorithmic management in the workplace (2025), recognised both the opportunities and the challenges associated with the increasing use of digitalisation, AI and algorithmic management in the workplace.

• Algorithmic management is most prevalent among clerical workers, technicians, and plant and machine operators. In these occupations, automated task allocation and performance monitoring are the most commonly reported algorithmic management functions, while automated scheduling is considerably less prevalent. Managers, professionals, services and sales workers, craft workers, workers in elementary occupations and especially skilled agricultural workers report lower levels of algorithmic management overall. Automated scheduling is also more evenly distributed across occupations than the other two functions.

• In the sectors with the highest prevalence of algorithmic management – finance and real estate, information and communications, and transport and storage – performance monitoring and automated task allocation are consistently the most frequently reported functions, whereas automated scheduling is less widespread. By contrast, all three functions are comparatively uncommon in education, agriculture and mining, and construction.

• Out of the three algorithmic management functions, automated task allocation is the strongest predictor of job transformation, including the creation of new tasks, the removal of other tasks and greater interaction with colleagues. Automated scheduling and performance monitoring show similar but more moderate associations with job transformation.

• Use of the three algorithmic management functions is associated with higher job demands, especially quantitative demands (capturing work intensity). Employees subject to algorithmic management are also more likely to experience job strain than those who are not. Among the three functions examined, performance monitoring stands out as the one most consistently associated with poorer health and well-being outcomes, and it appears to amplify, to a small extent, the adverse effects of existing job strain in already demanding jobs.

• Whether algorithmic management mitigates or compounds these negative outcomes depends on the specific algorithmic function: automated task allocation neither mitigates nor compounds job strain; automated scheduling has a marginally significant mitigating effect for employees in strained jobs; and performance monitoring significantly compounds the negative effects of job strain, further increasing exhaustion and work–family conflict and reducing work engagement.

• Algorithmic management is not evenly distributed across sectors, occupations and EU Member States, nor is it a uniform set of practices; rather, it encompasses a diverse range of managerial functions and organisational applications. While common baseline protections are essential, their implementation can be supported through social dialogue, collective bargaining, practical guidance and examples of good practice that reflect the different ways algorithmic management is used.

• Ongoing monitoring and comparisons across Member States can help strengthen evidence-based policymaking. At the same time, greater investment is needed in analysing company practices to better understand how algorithmic management functions are implemented and integrated into work organisation, including how different tools and practices interact and combine to shape workplace outcomes.

• Algorithmic management is part of the broader digital transformation of work. In this analysis, it is associated with higher job demands, especially quantitative demands and greater job strain, while its relationship with access to job resources is more mixed. As its adoption expands, attention should be paid to the implications for job quality by ensuring that its implementation is accompanied by adequate job resources, employee support and safeguards for healthy working conditions.

• Performance monitoring is the most widespread form of algorithmic management and is linked to higher demands, fewer resources and poorer health and well-being outcomes, especially for already strained workers. Safeguards should ensure transparency about what is monitored, how data are used and how decisions are made. Strong enforcement of labour and data protection laws – alongside the effective implementation of existing rights to information, consultation and collective bargaining – is essential to ensure that algorithmic performance monitoring is fair and proportionate.

• It is essential to ensure the responsible implementation and integration of algorithmic management functions within organisational processes and digital systems through governance frameworks, impact assessments and employee involvement. Where relevant risks are identified, organisations should implement appropriate safeguards in line with existing labour, data protection and AI legislation. Algorithmic management can reshape tasks and work organisation, requiring a human-centred redesign of jobs to align productivity gains with employee well-being, supported by policy guidance and consultation mechanisms.