'Managing and Monitoring Mobile Service Workers via Smartphone App: A case study on worker monitoring, algorithmic management and software for "field service management"' by Wolfie Christl comments
Mobile service workers, whether they are technicians, homecare workers or cleaning personnel, are increasingly being managed and monitored through smartphone apps. Data that was previously unavailable to employers is now being recorded and evaluated. As these apps give workers instructions about which client to visit next, how to get there and which tasks to perform at the client site, they become algorithmic managers. In the background, powerful software systems for “field service management” help employers organize, coordinate and schedule client visits, work orders and tasks. They promise to optimize, streamline and automate task allocation and help dispatchers and managers supervise workers, monitor their location, assess work performance and identify undesired behavior.
This case study explores how employers can use software and smartphone apps to manage and monitor mobile service workers. It focuses on the potential implications for employees in Europe and makes two contributions. First, it summarizes survey-based research on how employers actually use these technologies and how workers are affected. Second, it examines software that is available on the market. To illustrate wider practices, it investigates Microsoft software for “field service management”, which is part of the company’s comprehensive “Dynamics 365” system. The investigation aims to identify, examine and document data practices that affect workers, based on a detailed analysis of technical documentation and other publicly available sources.
Microsoft’s “field service management” system provides extensive functionality for algorithmic management, performance control and behavioral monitoring:
Via mobile app, workers receive instructions about work orders, client destinations, travel routes and a list of predefined tasks to be performed at client sites. The required arrival times and expected durations of work orders and tasks serve as target times. Workers confirm via app when they travel to clients, complete tasks or take breaks. Tasks can include sub-tasks with step-by-step instructions. Consequently, the app structures, directs and micromanages work, aligning it with rigidly defined processes. As it constantly reminds workers of time constraints and deviations, it includes implicit mechanisms for performance and behavior control.
Employers can utilize behavioral monitoring and performance control to supervise and pressure workers. Dispatchers can see their real-time location and travel routes on a map. They can monitor the current degree of completion of work orders and tasks. For completed work, they can see how much time a worker actually spent on it in relation to the target time. The system can remind both dispatchers and workers of cost limits associated with the time spent on work orders. To keep contractually agreed response times low, it can show a timer to dispatchers that counts the minutes and seconds that have elapsed since the creation of a work order.
New work orders can result from client inquiries or machines sending error codes, or they are automatically generated on a recurring basis, for example, based on service agreements. Organizations can define standardized work orders that include specific service tasks, instructions and estimated durations. The specified durations serve as target times and are used to distribute and schedule actual work orders to workers. The system can automatically schedule and dispatch work orders, and, as such, automatically assign tasks to workers. It can generate schedules for all workers for entire days or calculate a new schedule every 30 minutes. To match work orders to workers, it considers workers’ availability, location, predicted travel times and skill profiles. Automated scheduling is based on customizable optimization goals. The “maximize productivity” goal leads to schedules with minimized travel and idle times. The system can optionally create schedules that require workers to travel to or from client sites outside their working hours. Besides fully automated scheduling, it can also semi-automatically recommend workers for particular work orders.
Microsoft offers to predict how long it will take to complete particular work based on past data on work activities and “AI” models. It outlines possible reasons for deviations between predicted and previously specified durations by suggesting, for example, that a particular client, region, weekday, task or worker will likely increase the time required to carry out the work. As such, it may accuse workers of being slower than expected. This functionality can help dispatchers “enhance their team’s performance”, according to Microsoft.
The system is designed to rate and rank workers according to a wide range of performance and behavior metrics over the previous year, including by the number of completed work orders, the time spent on them in relation to target times and the time spent travelling, on break and in an “idle” state, i.e. without an assignment. Managers can identify undesirable behavior by evaluating how often workers missed the target time or arrived late to a client. They can assess workers by how much revenue they generate and by how satisfied the customers are with their work, based on surveys sent to clients after work was completed. Group-level metrics, for example, on the average time spent completing certain types of work in relation to targets, can also create pressure to speed up work. Employers can use Microsoft’s “Power BI” system to create almost any type of report.
While GPS location tracking is optional, many features rely on it. Microsoft recommends recording a worker’s location every “60 to 300 seconds”. Clients can be offered access to the current location of the scheduled worker and their estimated arrival time. The system systematically exposes personal data on worker behavior to employers, who can, for example, view records about workers’ exact whereabouts over time. It provides access to enriched location records that indicate, based on geofencing, at which time workers have entered or exited certain client sites or other areas. The system also provides records about remote assistance calls, including their start and end times, and summarizes how much time each worker spent on those calls.
Despite concerns about the reliability of “generative AI”, Microsoft has rushed to put its CoPilot technology into many products, including field service management. CoPilot promises to summarize information, create draft work orders based on customer emails and draft email replies. While Microsoft advertises it as a means to accelerate work, dispatchers are told to review “AI-generated content” because it can be “incorrect”. The system can be integrated with Microsoft 365 and other Microsoft software. Dispatchers and workers can handle work orders directly from within Outlook and Teams, turning them into task management systems.
Employers can customize Microsoft’s field service technology and use it in more or less problematic ways. Some intrusive pre-built reports are not available for German customers. Employers can add custom workflows to manage and monitor different types of mobile work. Microsoft’s Home Health system extends its field service technology with functionality specific to homecare. A brief investigation of field service technology offered by the major German vendor SAP shows that its system also offers intrusive performance monitoring functionality.
Implications for workers. The review of survey-based research on the practical use of field service technology in Austria, Norway, the UK and the US shows that these systems can have significant negative implications for workers. Digital task documentation in the name of billing, quality management or workload balancing can quickly evolve into far-reaching algorithmic management via app. Task direction and monitoring via smartphone app generally leads to increased surveillance and digital control at work. Employers may intentionally misuse the data for purposes other than it was originally collected, including for making negative decisions about workers. Standardized processes, rigid performance targets and automated scheduling can accelerate and intensify work and undermine work discretion. Knowledge that once resided with workers now increasingly lies within technical systems
