Workers waste 7+ hours weekly as 'AI middleware' fixing disconnected systems

A new Workday report exposes a growing paradox in enterprise AI adoption: while 45% of UK workers say AI has accelerated their work, one in four spend upwards of seven hours per week acting as 'human middleware'—manually moving data between disconnected AI systems, reconciling conflicting outputs, and feeding context into prompts. This inefficiency effectively neutralizes the productivity gains AI was meant to deliver. The study found that 62% of employees spend at least half their working time translating and coordinating between siloed systems rather than creating value. Worse, 60% frequently experience 'busy but unproductive' days, compared to 43% globally.

The core issue, according to Workday VP Daniel Pell, is that 'too many employees are serving as the human middleware between disconnected AI systems.' Companies struggle to embed AI deeply due to high volumes of approvals (27%), uneven AI training and access (26%), rigid workflows (26%), and poor data quality (24%). Ironically, 90% of workers still report a stronger sense of progress and connection to goals—suggesting a disconnect between perception and actual efficiency. Pell advises that the firms seeing real AI value are those building it directly into the systems where people, data, and work converge, rather than layering tools on top of fragmented infrastructure.