Artificial intelligence has officially entered the mainstream workplace, with nearly three-quarters of workers using it regularly, according to a comprehensive new study. But the real business value remains concentrated among a select group of companies that have moved beyond simply deploying AI tools to completely redesigning how work gets done.

The findings come from Boston Consulting Group's third annual "AI at Work 2025: Momentum Builds, But Gaps Remain​" survey, which polled over 10,600 workers across 11 countries. The research reveals a stark divide between AI adoption and AI transformation—a gap that could determine which companies thrive in the age of artificial intelligence.

While overall AI usage has surged to 72% of workers using it regularly, the growth has hit a significant roadblock among frontline employees. Only 51% of frontline workers are regular users—a figure that has stagnated compared to previous years, raising concerns about equitable access to AI benefits across organizational hierarchies.

The global picture shows interesting regional variations. Countries in the Global South continue to lead adoption rates, with India reaching 92% regular usage and the Middle East at 87%. However, these high-adoption regions also report the greatest anxiety about job displacement, with workers expressing far more concern than the global average of 41% who worry their roles could disappear within a decade.

Perhaps most concerning for IT departments and executives is the rise of "shadow AI"—unauthorized tool usage that poses growing security risks. More than half of respondents (54%) admitted they would use AI tools even without official approval, with younger workers particularly likely to circumvent company restrictions.

This unauthorized usage highlights a critical gap in organizational AI strategy. While employees are eager to harness AI's power, many feel inadequately supported by their employers, with only 36% saying they feel properly trained and just 25% of frontline workers reporting adequate leadership guidance.

The Value Realization Gap

The study's most significant finding centers on what BCG calls the "transformation divide." While most companies have focused on tool deployment, those achieving real business value are taking a fundamentally different approach: completely reimagining workflows and investing heavily in human transformation.

"Companies cannot simply roll out GenAI tools and expect transformation," said Sylvain Duranton, Global Leader of BCG X and report co-author. "Our research shows the real returns come when businesses invest in upskilling their people, redesign how work gets done, and align leadership around AI strategy."

These high-performing organizations are seeing tangible results that go beyond basic automation. Their employees report saving more time during workdays, spending increased time on strategic tasks, producing higher-quality outputs, and expressing greater confidence that AI is improving both decision-making and job satisfaction.

Looking ahead, the research identifies AI agents—sophisticated digital assistants capable of independent task management—as the next frontier. Three-quarters of surveyed employees believe these tools will be vital for future success, yet only 13% report that agents are currently integrated into their workflows.

Even more telling, just one-third of workers understand how AI agents function, suggesting a significant education gap that organizations must address. The study notes that as familiarity with agents increases, worker anxiety decreases, with employees beginning to view these tools as collaborators rather than competitors.

BCG's research outlines four critical imperatives for organizations serious about AI transformation:

Training Investment: Companies must dramatically increase their commitment to AI education, with the study showing that workers receiving five or more hours of training—particularly in-person with coaching—are significantly more likely to become regular users.

Value Tracking: Organizations need robust metrics to measure AI's impact on productivity, quality, and employee satisfaction, moving beyond simple adoption rates.

Workforce Transformation: The most successful companies are proactively reshaping workflows and building comprehensive upskilling and reskilling programs to support their workforce through AI-driven changes.

Agent Experimentation: Companies should begin rigorous testing of AI agents through A/B testing to accelerate learning and track both impact and potential risks.

"Companies that reshape their workflows and invest in people are seeing superior results," said Vinciane Beauchene, Global Lead on Human x AI at BCG and report co-author. "But that transformation must be accompanied by a clear people strategy and development engine to boost adoption and tackle the impacts it will have on work, the worker and the workforce."

The full report is avilable here.