Why it matters: Teaching employees to use AI rather than replacing them could add between $4.8 trillion and $6.6 trillion to the U.S. economy by 2034, according to Pearson’s new research released at Davos.
The big picture: Pearson's study reveals companies are pouring billions into AI infrastructure but (so far) seeing minimal enterprise productivity gains beyond coding. The missing link is the "learning gap" preventing employees from effectively using AI tools.
AI reached one billion users in three years, but learning hasn't kept pace - 59% of the global workforce will need reskilling by 2030, according to the World Economic Forum. Workers report saving time, but companies aren't seeing meaningful productivity lifts or ROI.
What to do: Pearson's DEEP Learning Framework offers four steps:
- Diagnose task-level augmentation opportunities
- Embed learning into daily workflows
- Evaluate skills progress systematically
- Prioritize learning as strategic investment
Dive deeper: The report highlights leading organizations that are investing in learning:
- IBM analyzes millions of files to infer skills, then embeds them in “badges” to link learning with career development.
- Microsoft leverages an agentic sales coach to create interactive and personalized coaching for thousands of employees.
- ServiceNow runs “colabs,” peer-to-peer learning sessions allowing AI experts to share their knowledge.
- Mindstone blocks weekly calendar time for employees to experiment with AI.
Bottom line: “The gains generated by AI are really hard to measure with our traditional productivity metrics, and new innovation always has a time lag to adoption,” said Michael Osborne, Professor of Machine Learning at the University of Oxford. “In fact, we might even see a J-curve effect—a short-term suppression of productivity—as organizations try out pilots, see some walk-backs, and have to re-allocate their investments.”
