Microsoft: AI Without Limits Fuels the Rise of the Endless Workday

The latest Microsoft Work Trend Index report reveals the blurring of work-life boundaries, as employees find themselves working in the mornings, evenings, and on weekends. The report’s authors highlight the possibility that AI may only amplify current workloads.

This research is grounded in an analysis of the behavior of millions of Microsoft 365 users and a survey that gathered insights from over 31,000 employees across 31 countries.

According to the findings, email and messaging activity begins before 6 a.m. and extends past 8 p.m. On average, a worker receives 117 emails daily—most of which are glanced at in under a minute. Notably, engagement on Microsoft Teams surpasses that of email services by 8 a.m.

Moreover, over a third of employees respond to emails late in the evening, and 16% participate in online meetings after the official workday has concluded.

On weekends, the use of Word, Excel, and PowerPoint surpasses that of Teams and other applications, as employees feel they can finally carve out time for uninterrupted focused work.

As a result, nearly half of the workforce (48%) and more than half of managers (52%) express that their work feels chaotic.

One contributing factor to this phenomenon is the highly fragmented workday. Employees receive notifications every 2 minutes on average and switch between tasks up to 280 times per day. Over half of the meetings lack an agenda, and 10% are scheduled at the last minute.

Analysts from Microsoft contend that artificial intelligence, which could automate certain routine tasks, will only increase overload without changes to work structure. They point out that automation without defined priorities and new guidelines could merely «accelerate the treadmill effect.»

Consequently, the company advises a shift to a new working model that includes clear time periods for focusing on tasks, implementing «quiet hours» for teams, and training employees to work collaboratively with AI as a partner rather than simply a tool.

Meanwhile, an analysis by METR Evaluations indicates that AI systems’ ability to assist in task execution is improving rapidly. The duration of autonomous AI operations is doubling every seven months, and by 2027, these systems are expected to manage an eight-hour workday with a success rate of 50%.