In the wake of the Covid-19 pandemic, companies everywhere have been forced to make big changes as they adjust to the new realities of employees working from home. Given how dependent remote work is on digital technology, it is not surprising that many business leaders have chosen to apply data-driven decision-making to this process. Unfortunately, though, a few companies have taken this technical approach to business administration a bit too literally. Case in point is Xsolla, a Russia-based payment processing company that recently used data analytics to justify the firing of 147 employees. In a leaked email sent out to terminated workers, CEO Alexandr Agapitov explained things as follows: "My big data team analyzed your activities in Jira, Confluence, Gmail, chats, documents, dashboards and identified you as an unengaged and unproductive employee." He continued, "...you were not always present at the workplace when you were working remotely." At the time of writing, just about a month has gone by since the dismissals, and Agapitov's decision has shown itself to be, to say the least, shortsighted. The company is currently facing a social media firestorm that is proving hard to put out. More pressingly, Xsolla must now devote an enormous amount of resources to recruiting and training new employees. IT is an extremely competitive labor market, and the company hasn't made things easy for itself with this PR fiasco. So what should the company have done instead? Well today, there exist a plethora of ways that companies can harness AI and machine learning algorithms to optimize workforces - including the three outlined below. By applying just one of these strategies, Xsolla could have avoided their firing scandal entirely. Smart Scheduling Low productivity at Xsolla was more likely than not the result of poor time management. In other words, employees were poorly equipped to set up and manage their own schedules while working from home. Indeed, this has been a common tendency throughout the workforce, across industries. Moreover, it seems that the management staff at Xsolla did little to help workers learn effective time management skills. Fortunately, algorithms can be easily developed to determine the hours during which individuals work best. Only a few months of cloud-based work is enough to gather data and determine when, over the course of a day, the most value is created by any particular employee. With this data at hand, it only takes a few lines of code to generate each employee's ideal working schedule. This goes a lot of the way towards automating time management. To illustrate, if an engineer is shown to mark the most number of tasks completed in the morning, he could be issued a 40-hour - AI-generated - schedule that focuses on this time of day. A project manager's job, in this case, would be to encourage the engineer to stick to this schedule and to provide input as to when they feel they do their best work. As the algorithm gradually collects more data, including from employee feedback, the schedule can be enhanced to perfectly fit the worker's lifestyle and productivity patterns. Within a year of working this way, the employee will likely have gained the habit of working productively 40 hours a week anyway, and the crutch of the AI-generated schedule may not even be needed any longer. Automated Reminders As this smart scheduling system continues to optimize each employee's ideal working regimen with time, it will eventually become possible to apply machine learning in order to make predictions about when team members will work productively in the future. Then, a simple mobile application can be developed to send push notifications - phrased as words of encouragement - to employees whenever the algorithm thinks the most work can be done. Any well-intentioned team member would be more than happy to get back to work upon receiving a message like "Work now for an hour and get your task done in half the time!" Of course, these notifications should be programmed to turn themselves off automatically following the completion of each day's tasks. Overwork is just as bad as underwork, after all! Targeted Accommodation The truth of the matter is that not everyone is well accommodated for remote work. Many workers face external factors that keep them from doing good work at home. Data analysis can be used to identify these issues within a team and to help find solutions. For example, an algorithm can be written to check across the team's activities to detect who is frequently logging in from different IPs. The company can understand from this data that the individual may not have a single workspace available to them. Maybe they live in a noisy area and so split their time between different locations, or perhaps they lack a good internet connection at home and are forced to hop from café to café. To solve the problem, a meeting could be set up with HR. In the first case, the company could help to find a good, secluded, location to work. Second, the IT department could assist in setting up a stable internet connection in the employee's home. Conclusion With many workers expecting to work from home at least one day a week in the future, there is certainly a lot of room for workforce optimization. In fact, the above implementations are just a few ways companies can harness emerging technologies like big data analytics, AI, and machine learning to increase productivity for remote workers. So, if one thing's for sure, it's that Xsolla missed out on a great opportunity to innovate. Better luck next time!
