The AI at Work Index 2026 finds that 88% of organizations use AI in at least one business function, but only 5% report transformative returns, because employees in more than 90% of organizations do much of their AI work through personal ChatGPT and Claude accounts that the enterprise does not control. 5W's study covers ten professions and five countries and synthesizes more than 80 authoritative sources.
Read the full AI at Work Index 2026.
How many organizations use AI, and what do they get from it?
88% of organizations now use AI in at least one business function, and only 5% report transformative returns. In more than 90% of organizations, employees use personal AI tools for work, often many times per day, through accounts the enterprise does not control and the board does not see.
MIT Project NANDA identified the split directly: only 40% of companies have purchased official LLM subscriptions, while workers at more than 90% of the companies surveyed reported regular use of personal ChatGPT, Claude, Gemini and Copilot for work. KPMG found that 50% of US workers use AI at work without knowing whether it is allowed, and 44% knowingly use it improperly. Harmonic Security's 2026 analysis of 22 million enterprise AI prompts tracked traffic across 665 different AI tools, most of them unsanctioned.
This is the shadow AI economy: AI use that happens outside company-approved tools. It is where much of the real work is happening, and it is also where undocumented compliance risk sits.
What are the six findings of the 2026 Index?
The Index reports six findings that together define the 2026 picture of workplace AI.
1. The shadow AI economy is the productivity reality. Enterprise spend is not landing on the P&L, while personal-account use is. Organizations that channel the shadow instead of suppressing it capture both the productivity and the governance.
2. The leader and employee gap is large and consequential. Microsoft's 2025 Work Trend Index surveyed 31,000 workers across 31 countries: 67% of leaders are familiar with AI agents versus 40% of employees, and 79% of leaders believe AI will accelerate their careers versus 67% of employees. McKinsey found that C-suite leaders are more than twice as likely to blame employee readiness for slow adoption as they are to blame their own role.
3. Profession predicts AI use more than any other variable. The function gap is bigger than the country gap, as the table shows.
| Profession | AI adoption | Source |
|---|---|---|
| Software developers | 84% adoption, 51% daily | Stack Overflow 2025 |
| Legal | 79% | Clio 2025 |
| Marketing | Approximately 72% generative AI use | Jasper, State of AI in Marketing 2025 |
| Frontline white-collar | 51%, stalled | BCG 2025 |
4. Country variance is real, and mostly a proxy for profession mix. India's 92% weekly use reflects a heavy concentration of IT services, and Japan's 51% reflects its manufacturing weight. For US multinationals, country benchmarks suit positioning, and profession benchmarks suit program design.
5. Training is the largest unforced variable. BCG found that when employees receive strong leadership support and at least five hours of training, positive sentiment toward AI rises from 15% to 55%. KPMG found that 83% of AI-using employees say they need to improve their skills, and fewer than half say their employer provides sufficient training. Stack Overflow reports that developer sentiment toward AI tools fell from above 70% in 2023 and 2024 to 60% in 2025 even as usage climbed, so trust is eroding among the most sophisticated users.
6. The employer's playbook has six moves. Measure six things: frequency, depth, task diversity, acceptance rate, the sanctioned versus shadow split, and productivity perception. Close the leader and employee gap deliberately, channel the shadow instead of banning it, invest in role-specific training, build governance clarity before scaling access, and communicate consistently.
Why is AI at work a communications problem?
AI at work is a communications problem because every driver of adoption, including leadership modeling, governance clarity, training visibility and the move from shadow to sanctioned tools, depends on how it is communicated.
"The productivity already exists inside your company. It is sitting in personal ChatGPT accounts that your IT department does not control and your board does not see. The job in 2026 is not to prove AI works. It is to bring the work that already works out of the shadows and into the business."
Ronn Torossian, Founder and Chairman, 5W AI Communications
The companies that come out ahead will not necessarily be the ones with the biggest AI budgets. They will be the ones whose message, internal to employees, external to buyers and upward to boards, closes the gap between the productivity that already exists and the enterprise value it should create.
How was the AI at Work Index built?
The AI at Work Index is a synthesis study drawing on more than 80 authoritative sources. These include the Microsoft 2025 Work Trend Index, McKinsey's State of AI 2025, MIT Project NANDA's State of AI in Business 2025, BCG's AI at Work 2025, the KPMG and University of Melbourne Trust in AI Global Study 2025, Stack Overflow's 2025 Developer Survey, Wharton Human-AI Research, PwC's 2025 AI Agent Survey, the Clio 2025 Legal Trends Report and Jasper's State of AI in Marketing 2025, plus material from Stanford HAI, the OECD and the American Bar Association.
5W organizes the findings through a framework of six metrics of adoption and six drivers of adoption depth.
CONCLUSION
The AI at Work Index 2026 shows that the productivity from AI is real and is running outside the enterprise.
For research licensing, custom engagements or media inquiries, contact research@5wpr.com. More commentary from Ronn is at ronntorossian.com.
Frequently asked questions
What is the AI at Work Index 2026?
The AI at Work Index 2026 is a 5W study of workplace AI across ten professions and five countries. It synthesizes more than 80 authoritative sources into six findings and a six-move employer playbook.
What is shadow AI?
Shadow AI is employee use of personal AI accounts, such as ChatGPT or Claude, for work outside company-approved tools. The Index reports that employees at more than 90% of organizations do this.
How many organizations see transformative returns from AI?
Only 5% report transformative returns, even though 88% of organizations use AI in at least one business function.
Why does training matter for AI adoption?
BCG found that strong leadership support plus at least five hours of training raised positive sentiment toward AI from 15% to 55%. KPMG found that fewer than half of AI-using employees say their employer provides sufficient training.
What should employers do about shadow AI?
The Index recommends channeling shadow use instead of banning it. That means measuring adoption, closing the leader and employee gap, investing in role-specific training, setting governance rules before scaling access, and communicating consistently.
Ronn Torossian is the founder and chairman of 5W AI Communications. He is the publisher of Everything-PR and the author of two best-selling editions of For Immediate Release.
