Corporate America has spent an estimated $50 billion on AI tools in the past year, with executives claiming transformative productivity gains. But a new study from Foxit Software suggests workers are experiencing something very different: an average time savings of just 16 minutes per week.
Let's do the math. Sixteen minutes per week equals roughly 14 hours annually per employee. For a company with 10,000 workers earning an average of $75,000, that's about $5 million in theoretical productivity gains—assuming those 16 minutes are spent on revenue-generating activities, which is optimistic.
Now compare that to the cost side. Microsoft 365 Copilot runs $30 per user per month. For our hypothetical 10,000-person company, that's $3.6 million annually. Salesforce's Einstein AI? Similar pricing. Add in implementation costs, training, IT overhead, and the inevitable productivity dip during rollout, and the ROI calculation starts looking grim.
The Foxit study, which surveyed 1,200 knowledge workers across multiple industries, found a striking disconnect. 76% of executives reported that AI had significantly improved their teams' productivity. But when researchers asked the workers themselves, the story changed. Most reported minor time savings on routine tasks like email summarization and document formatting—useful, but hardly transformative.
More concerning: 43% of workers said AI tools had actually increased their workload, primarily due to the need to verify AI-generated outputs, correct errors, and manage the integration of multiple AI systems that don't always play nicely together.
This is reminiscent of the 1980s productivity paradox, when massive investments in personal computers failed to show up in economic data. Nobel laureate Robert Solow famously quipped, "You can see the computer age everywhere but in the productivity statistics." It took nearly a decade for organizations to reorganize workflows and train workers before PC investments paid off.
The AI productivity story may follow a similar trajectory. The technology is real, but the organizational change required to fully leverage it is slow, expensive, and disruptive. Companies are essentially paying for the promise of future productivity while experiencing the pain of present-day transition costs.
