Future of work

Note: The research presented in this paper is based on surveys and interviews from 300 organizations during the first quarter of 2025 and reflects reported adoption patterns rather than measured productivity outcomes. The 90% employee usage figure represents self-reported regular use of AI tools for work tasks and may include varying levels of sophistication. 

Employees are crossing the GenAI adoption gap on their own, even as their employers remain stuck in pilot phases. In a study examining AI implementation at 300 enterprises, workers from more than 90% of the surveyed companies reported regularly using personal AI tools for work tasks—while only 40% of those same companies have purchased official AI subscriptions for their workforces.  

This shadow AI economy demonstrates that workers are already integrating AI into their daily workflows through consumer tools like ChatGPT and Claude, usually without IT oversight or formal training. Workers report using these personal AI tools multiple times daily even as their companies’ official AI initiatives remain stalled in experimental phases. The gap between individual adoption and organizational deployment reveals that workers are not waiting for formal systems to catch up.  

The preference for consumer AI tools over enterprise solutions is due to usability differences. Workers consistently praise consumer tools for their flexibility, familiarity, and immediate utility while describing enterprise AI tools as rigid, overengineered, or misaligned with workflows. One surveyed corporate lawyer reported having invested in a specialized contract analysis tool but consistently defaulting to ChatGPT for drafting work, noting that the latter provides better output quality and allows iterative refinement through conversation. This pattern suggests that general-purpose tools that offer conversational interfaces often outperform specialized enterprise systems that cost significantly more.  

This trend toward informal adoption has significant  implications for workforces. Workers who actively use AI tools personally report productivity gains and develop practical AI literacy through hands-on experience while their companies’ formal AI initiatives stall. This creates knowledge asymmetry whereby frontline employees may understand AI capabilities better than the organization’s leaders. The shadow economy also raises questions about equitable access because workers who can afford consumer AI subscriptions or have technical aptitude gain advantages over colleagues without these resources or skills.  

Why these findings matter

For the public sector: The disconnect between organizational AI strategy and real-world worker adoption patterns suggests that workforce AI policies should account for the development of informal skills that is already occurring. Training programs and digital literacy initiatives need to recognize and build upon the practical AI experience that workers are gaining through consumer tools rather than starting from assumptions that the workforce lacks AI exposure.  

For employers and employment services: The shadow AI economy indicates that workers are ready and willing to integrate AI into their work, but their organizations are not keeping pace. Employers might benefit from understanding which personal AI tools workers already use and why they often prefer these over costly enterprise alternatives. This knowledge could inform procurement decisions and implementation strategies that align with actual workflow needs rather than following top-down technology mandates.  

For post-secondary institutions and students: The research suggests that practical, hands-on AI experience may matter more than formal credentials for near-term workforce readiness. Students already use consumer AI tools to enhance their coursework and projects, and may be developing relevant workplace skills in doing so, even if these tools differ from those used by enterprise systems. Institutions should consider how to validate and build upon this informal learning.  

For people in the LMI ecosystem: The shadow AI economy represents a significant measurement challenge for labour market data. Traditional metrics of technology adoption and digital skills may miss the substantial AI capability development that is happening outside formal channels. LMI professionals should consider how to capture data on the informal use of AI tools and how this use relates to productivity and job-quality outcomes.