Great tech rebellion? 75% workers bypassing company software: Report

A new report reveals that the modern workplace has fallen into a "workaround economy," where employees are increasingly bypassing official company software to piece together their own tech stacks, driven by fragmented systems and mounting cognitive overload. According to the study conducted by The Harris Poll on behalf of AI productivity platform Superhuman, 75% of professionals now operate outside of fully approved tool stacks, relying on a mix of company-provided and personally selected applications to get their jobs done. The survey of 1,505 professionals highlights a growing crisis in enterprise technology: while workers remain optimistic about their tools, the friction of moving work across disconnected systems is forcing them into unofficial, often insecure, habits.

The data shows that the burden of these disconnected systems is leading to widespread "shadow IT" and a heavy mental toll. Nearly half of workers, 48%, report that at least half of their tools are personally selected rather than provided by their employer. More alarming for enterprise security is the finding that 65% of workers are actively seeking out unofficial AI tools, bypassing corporate governance. Furthermore, 61% of employees admit to preferring familiar tools even when better, company-approved alternatives exist, simply to reduce the mental friction of navigating new systems. 

The root of the problem, the report argues, is not "tool fatigue," but rather a massive increase in cognitive load. Productivity rarely breaks down during deep, focused work; instead, it fractures in the seams between tasks. The survey found that the biggest productivity drops occur when coordinating with others and switching between tasks, both cited by 32% of workers, followed by gathering information and tracking progress at 26% and 25%, respectively. In high-cognitive-load environments, the mental tax is severe: 96% of workers say that deciding how to do their work takes as much effort as actually doing it, compared to just 43% in low-load environments. Similarly, 95% of those in high-load settings report that important information is spread across too many systems, compared to only 18% in low-load environments.

Rather than solving these workflow fractures, artificial intelligence is often amplifying them. The report emphasizes that AI inherits the environment it enters, and in highly fragmented workplaces, it exposes complexity instead of eliminating it. Among workers in high-load environments, 91% say AI tools add more options than clarity, 88% feel AI is designed around features rather than how they actually work, and 83% say AI creates extra work for them to review or fix. The value of AI also visibly dilutes when tasks are scattered. For example, while 69% of workers using a single email tool say AI reduces their effort, that number drops to 57% for those juggling three or more email applications. 

Ultimately, the findings suggest that the next major gains in enterprise ROI will not come from purchasing more software or adding new AI features, but from designing systems that actually align with human workflows. By forcing employees to act as the integration layer between disconnected apps, companies are suffering from fragmented data, duplicated tools, and diminished returns on their technology investments. Closing what the report calls the "last mile of AI" requires leaders to prioritize platforms that carry context across systems and eliminate the need for workers to constantly bridge the gaps on their own.