AI at Work statistics
63 statistics · 21 primary sources · data as of November 2026
Every figure below links to the institution that published it. Each was opened and checked by a human editor.
Section 01
46% of global leaders said their companies are using AI agents to fully automate workflows or processes in 2025
Microsoft (2025 Work Trend Index) · 2025 · Same study; automation concentrated in customer service, marketing and product development · source
81% of leaders expected AI agents to be moderately or extensively integrated into their company's AI strategy within the next 12–18 months (2025)
Microsoft (2025 Work Trend Index) · 2025 · Same study; expectation measure, not current deployment · source
46% of U.S. employees used AI at work at least a few times a year in Q4 2025, with 26% using it at least weekly and 12% daily
Gallup · 2025 · Gallup Panel survey of 22,368 employed U.S. adults, Oct 30–Nov 14, 2025; self-reported individual usage (different 口径 from firm-level surveys) · source
49% of U.S. employees said they never use AI in their role in Q4 2025
Gallup · 2025 · Same Gallup Q4 2025 survey; non-use share roughly flat even as frequent use rises · source
About 1 in 5 U.S. businesses (19.8%) used AI in some business function as of May 2026, with usage hovering between 17% and 20% from December 2025 to May 2026
U.S. Census Bureau (Business Trends and Outlook Survey) · 2026 · BTOS biweekly nationally representative survey of ~1.2M businesses; AI question revised Nov 2025 to cover 'any business function', so not comparable with the narrower pre-2026 'producing goods or services' series; firm-level measure · source
37% of U.S. firms with at least 250 employees reported using AI in 2026, vs under 20% of firms with four or fewer employees
U.S. Census Bureau (Business Trends and Outlook Survey) · 2026 · Same BTOS window (Dec 14, 2025–May 3, 2026); Information sector 39.7% and Finance/Insurance 33.9% vs Retail ~14% · source
Organizational AI adoption worldwide reached 88% of surveyed organizations in 2025, up from 55% in 2023; gen AI is used in at least one business function at 70%, while AI agent deployment stayed in single digits
Stanford HAI (AI Index Report 2026) · 2026 · Synthesis of McKinsey Global Survey data (org-reported); Europe 91%, North America 90%; agent use measured separately and still early · source
Companies expanded worker access to sanctioned AI tools from under 40% to around 60% of workers in one year (2025); among workers with access, fewer than 60% use AI in their daily workflow
Deloitte (The State of AI in the Enterprise) · 2026 · Annual survey of 3,235 director-to-C-suite leaders across 6 industries and 24 countries, Aug–Sep 2025; access measure ≠ active usage · source
Section 02
Field studies report AI productivity gains of 14–15% in customer support work
Stanford HAI (AI Index Report 2026) · 2026 · Synthesis of peer-reviewed field studies; gains largest in structured, measurable work where outputs are easy to monitor · source
Field studies report AI productivity gains of 26% in software development work
Stanford HAI (AI Index Report 2026) · 2026 · Same AI Index synthesis; gains smaller in tasks requiring deeper reasoning · source
Field studies report AI gains of 50% in marketing output volume
Stanford HAI (AI Index Report 2026) · 2026 · Same AI Index synthesis; report also flags possible long-term learning penalties from heavy AI reliance · source
Access to a generative AI assistant increased customer support agents' productivity by 14% on average, with a 34% improvement for novice and low-skilled workers
Brynjolfsson, Li & Raymond — Quarterly Journal of Economics (NBER WP 31161) · 2025 · Field data from 5,179 customer support agents; productivity = issues resolved per hour; published QJE 140(2):889–942; effect smallest for experienced/high-skilled agents · source
In a preregistered experiment, professionals using ChatGPT completed mid-level professional writing tasks 40% faster with 18% higher output quality
Noy & Zhang — Science · 2023 · Randomized online experiment, 453 college-educated professionals, occupation-specific incentivized writing tasks; quality rated +18%; inequality between workers decreased · source
Employees using AI reported an average 40% boost in productivity in 2025
Upwork Research Institute · 2025 · Global survey of 2,500 workers (1,250 C-suite, 625 full-time employees, 625 freelancers), conducted Mar 25–Apr 9, 2025; self-reported, not measured output · source
77% of employees reported in 2024 that AI had increased their workload
Upwork Research Institute · 2024 · Prior-year (2024) study finding cited in the July 2025 'From Tools to Teammates' release; self-reported; counterweight to headline productivity gains · source
Among U.S. workers who have used AI chatbots for work, 40% said they were extremely or very helpful for doing things more quickly, vs 29% for improving work quality
Pew Research Center · 2025 · Survey of 5,273 employed U.S. adults, Oct 7–13, 2024; self-reported helpfulness among chatbot users · source
Section 03
Goldman Sachs Research estimates about 300 million jobs globally are exposed to automation by AI
Goldman Sachs Research · 2026 · Model-based estimate extrapolating U.S./Europe occupational task-content data to employment-weighted global totals; 'exposed to automation' ≠ 'displaced' — most exposed jobs are only partially automatable · source
AI can potentially automate tasks that account for 25% of all U.S. work hours
Goldman Sachs Research · 2026 · Task-content analysis of 900+ occupations; task-hours measure, not a job-loss count · source
Goldman Sachs' base case sees 6–7% of U.S. workers displaced during a roughly 10-year AI adoption transition, adding about 0.6 percentage points to the unemployment rate
Goldman Sachs Research · 2026 · Projection under a ~10-year diffusion timeline; a faster, frontloaded transition would have larger economy-wide impacts · source
U.S. employment of software developers aged 22–25 fell nearly 20% from 2024 to 2025
Stanford HAI (AI Index Report 2026) · 2026 · AI Index labor-market analysis (LinkedIn/Lightcast data); effect concentrated in the youngest cohort of an AI-exposed occupation; older age groups grew · source
Among U.S. workers aged 22–25, employment in the most AI-exposed occupations fell roughly 16% relative to the least-exposed occupations after controlling for firm-type effects
Stanford HAI (AI Index Report 2026) · 2026 · Regression-adjusted estimate isolating AI exposure from broader shocks (rates, sector slowdowns); gap began widening mid-2024; relative measure, not an absolute employment level · source
32% of surveyed organizations expected AI to reduce their workforce over the coming year, while 43% expected little or no change
Stanford HAI (AI Index Report 2026, citing McKinsey 2025 survey) · 2026 · Employer expectation measure; anticipated reductions highest in service operations, supply chain and software engineering; large-scale job losses not yet visible in aggregate employment data · source
In the U.S., entry-level hiring in AI-augmented occupations (e.g., software engineer, data analyst) fell 8.9% year-over-year in early 2026, vs a 1.9% decline in overall hiring
LinkedIn Economic Graph (AI Labor Market Update) · 2026 · LinkedIn hiring data across US/UK/France/Germany/India, last-three-months window; 1.3B+ member base; hiring-demand measure, not employment levels · source
U.S. startups now operate with 26% fewer employees in their first year compared with 2021, a shift LinkedIn links in part to AI-driven efficiencies
LinkedIn Economic Graph (AI Labor Market Update) · 2026 · LinkedIn company headcount analysis; correlation with AI adoption inferred, not experimentally identified · source
The World Economic Forum projects 170 million new jobs created and 92 million displaced by 2030 — net growth of 78 million, a 22% churn of today's formal employment
World Economic Forum (Future of Jobs Report 2025) · 2025 · Projection extrapolating a survey of 1,000+ employers (55 economies, 14M workers covered) with ILO global employment data; driven by macrotrends beyond AI (demographics, green transition, geoeconomics) · source
About a third (33%) of respondents in McKinsey's 2025 survey expected their organization's workforce to decline in size because of AI
McKinsey & Company · 2025 · The State of AI in 2025 survey, N=1,993, Jun 25–Jul 29, 2025; expectation measure — a small share also expected headcount to increase · source
AI skills appeared in 2.5% of all U.S. job postings in 2025, up 55% year-over-year
Lightcast (data partner, Stanford AI Index 2026) · 2026 · Analysis of billions of U.S. job postings; labor-demand measure; +72% vs 2022, +297% vs a decade ago · source
Section 04
Global corporate AI investment more than doubled to a record $581.69 billion in 2025
Stanford HAI (AI Index Report 2026) · 2026 · Total = private investment ($344.66B, +127.5% YoY) + M&A ($214.44B); private investment now ~60% of the total · source
Private generative AI investment grew more than 200% in 2025 and captured nearly half of all private AI funding
Stanford HAI (AI Index Report 2026) · 2026 · Gen AI led the private-investment surge; newly funded AI companies rose 71% and billion-dollar funding events nearly doubled · source
U.S. private AI investment reached $285.88 billion in 2025, 23 times China's $12.41 billion
Stanford HAI (AI Index Report 2026) · 2026 · Cross-country comparison; report notes China's state guidance funds (est. $184B deployed 2000–2023) understate its total AI spending · source
Google reported more than $150 billion in annual capital expenditures in 2025 as major cloud providers accelerated AI infrastructure spending
Stanford HAI (AI Index Report 2026) · 2026 · Frontier AI compute costs also hit records: OpenAI $16.3B and Anthropic $6.8B compute spend in 2025 (R&D + inference) · source
Despite $30–40 billion in enterprise generative AI investment, MIT NANDA estimates 95% of organizations are getting zero return, with just 5% of integrated AI pilots extracting millions in value
MIT NANDA (The GenAI Divide: State of AI in Business 2025) · 2025 · Multi-method study: 300+ public AI implementations analyzed, 52 structured interviews, 153 senior-leader survey responses; returns measured as measurable P&L impact · source
In MIT NANDA's data, 60% of organizations evaluated GenAI tools, but only 20% reached the pilot stage and just 5% reached production
MIT NANDA (The GenAI Divide: State of AI in Business 2025) · 2025 · Same study; failures attributed to brittle workflows and lack of contextual learning rather than model quality · source
Organizations that bought external AI tools rather than building in-house saw roughly twice (2×) the success rate, per MIT NANDA's analysis
MIT NANDA (The GenAI Divide: State of AI in Business 2025) · 2025 · Same study's 'implementation advantage' pattern; successful adopters also empowered line managers over central labs and chose tools that integrate deeply · source
Only 25% of organizations have moved 40% or more of their AI experiments into production to date, though 54% expect to reach that level within three to six months (2025 survey)
Deloitte (The State of AI in the Enterprise) · 2026 · N=3,235 director-to-C-suite leaders, 24 countries, Aug–Sep 2025; self-estimated experiment-to-production conversion · source
74% of companies plan to deploy agentic AI within two years, but only 21% report having a mature governance model for autonomous agents
Deloitte (The State of AI in the Enterprise) · 2026 · Same 2025 survey; oversight lagging deployment intent · source
78% of leaders were considering hiring for AI-specific roles in 2025, rising to 95% at Frontier Firms
Microsoft (2025 Work Trend Index) · 2025 · Roles under consideration include AI trainers, data specialists, security specialists, AI agent specialists, ROI analysts and AI strategists; consideration measure, not filled positions · source
In 2025, 45% of leaders ranked expanding team capacity with digital labor as a top priority, 47% ranked upskilling their existing workforce first, and a third were considering headcount reductions
Microsoft (2025 Work Trend Index) · 2025 · Same study; ranked workforce strategies for the next 12–18 months · source
78% of organizations expected to increase overall AI spending in the next fiscal year, even as regulatory compliance became the top deployment barrier (38%, up from 28% in wave one)
Deloitte (State of Generative AI in the Enterprise, Q4 report) · 2025 · Wave-four survey of 2,773 director-to-C-suite leaders across 14 countries, conducted Jul–Sep 2024; nearly three-quarters also said their most advanced GenAI initiative is meeting or exceeding ROI expectations · source
Section 05
52% of U.S. workers felt worried about how AI may be used in the workplace in the future, vs 36% hopeful and 33% overwhelmed
Pew Research Center · 2025 · Survey of 5,273 employed U.S. adults, Oct 7–13, 2024; emotions about future workplace AI use (not current usage) · source
Only 6% of U.S. workers said workplace AI will lead to more job opportunities for them in the long run, while 32% expected fewer and 31% expected no difference
Pew Research Center · 2025 · Same Pew survey (N=5,273, Oct 2024); lower- and middle-income workers more likely to expect fewer opportunities · source
16% of U.S. workers said at least some of their work is done with AI, while 63% said they don't use AI much or at all in their job (2024 data)
Pew Research Center · 2025 · Same Pew survey, fielded Oct 7–13, 2024; only 2% said all or most of their work is done with AI; AI users skew young (73% under 50) and college-educated (51%) · source
79% of leaders believed AI will accelerate their careers, vs 67% of employees (2025)
Microsoft (2025 Work Trend Index) · 2025 · 31-country study; leaders ahead of employees on all seven 'agent boss mindset' indicators measured · source
43% of workers globally feared AI could take their jobs, vs only 21% of Frontier Firm leaders (2025)
Microsoft (2025 Work Trend Index) · 2025 · Same study; fear gap between Frontier Firm leaders and the global workforce · source
80% of the global workforce said they lack enough time or energy to do their work, while 53% of leaders said productivity must increase (2025)
Microsoft (2025 Work Trend Index) · 2025 · Same study; 'capacity gap' combining survey and Microsoft 365 telemetry (employees interrupted every ~2 minutes during core hours, 275 interruptions/day) · source
Among U.S. employees who don't use AI, 44% said the main reason is they don't believe AI can assist with their work, while only 16% cited lack of access to AI tools
Gallup · 2025 · Gallup workforce survey Q2 2025, N=19,043 employed U.S. adults; perceived utility, not access, is the dominant stated barrier · source
88% of freelancers said AI has positively impacted their careers, and 90% said it helps them learn faster (Upwork, 2025)
Upwork Research Institute · 2025 · Global survey of 2,500 workers (incl. 625 freelancers), Mar 25–Apr 9, 2025; freelancers report stronger AI-linked well-being than full-time peers · source
42% of freelancers credited AI with helping them specialize in a particular niche (2025)
Upwork Research Institute · 2025 · Same Upwork study; freelancers primarily use AI as a learning partner · source
71% of Frontier Firm leaders said their company is thriving, vs just 39% of workers globally (2025)
Microsoft (2025 Work Trend Index) · 2025 · Same study; Frontier Firms = 844 of 9,037 surveyed leaders meeting org-wide deployment + ROI criteria; perception gap between leaders and workforce · source
Section 06
Employers estimate 59% of workers will need training by 2030 — 29 upskilled in current roles, 19 reskilled and redeployed, and 11 unlikely to receive it (WEF)
World Economic Forum (Future of Jobs Report 2025) · 2025 · Future of Jobs Survey: 1,000+ employers representing 14M+ workers across 55 economies; employer-projected training need through 2030 · source
Employers expect 39% of workers' core skills to change by 2030, down from 44% in the 2023 edition
World Economic Forum (Future of Jobs Report 2025) · 2025 · Same survey; 'skill instability' stabilizing at a high level after COVID-era peaks (57% in 2020) · source
50% of the global workforce had completed training as part of employers' learning and development initiatives in 2025, up from 41% in 2023
World Economic Forum (Future of Jobs Report 2025) · 2025 · Same survey; training-completion share rose across nearly all industries · source
77% of employers plan to upskill their existing workforce to work more effectively alongside AI by 2030, and AI & big data ranks as the fastest-growing skill category
World Economic Forum (Future of Jobs Report 2025) · 2025 · Same survey; most-anticipated workforce strategy in 45 of 55 economies covered · source
63% of employers identified skills gaps as a major barrier to business transformation over 2025–2030, and 85% plan to prioritize upskilling their workforce
World Economic Forum (Future of Jobs Report 2025) · 2025 · Same survey; skill gaps are the single biggest transformation barrier cited · source
54% of skilled freelancers reported advanced or expert-level AI tool proficiency in 2025, vs 38% of full-time employees; 51% of freelancers self-train on AI tools vs 39% of FTEs
Upwork Research Institute (Future Workforce Index 2025) · 2025 · Survey of 3,000 U.S. skilled knowledge workers, Dec 2024–Feb 2025; freelancers also use AI more often (62% several times a week vs 53% of FTEs) · source
67% of leaders said they are familiar or extremely familiar with AI agents, vs 40% of employees (2025)
Microsoft (2025 Work Trend Index) · 2025 · Same study; familiarity gap cited as an upskilling urgency signal (leaders 67%, employees 40%) · source
Companies can grow their AI talent pipeline 8.2x globally by adopting skills-based hiring instead of screening on degrees or job titles
LinkedIn Economic Graph (The Skills Signal Report 2025) · 2025 · LinkedIn Economic Graph modeling; expansion factor for AI talent pipeline under skills-first hiring · source
84% of companies have not redesigned jobs or the nature of work around AI capabilities, and insufficient worker skills is seen as the biggest barrier to AI integration
Deloitte (The State of AI in the Enterprise) · 2026 · N=3,235 leaders, 24 countries, Aug–Sep 2025; most firms focus on employee AI education rather than rearchitecting roles and workflows · source
Mentions of 'Agentic AI' skills in U.S. job postings rose over 280% in one year, from 0.06% of postings in 2024 to 0.23% in 2025 (~90,000 postings)
Lightcast (Stanford AI Index 2026 data partner) · 2026 · New Agentic AI skill cluster tracked in 2025 (Lightcast now tracks 300+ AI skills in 10 clusters); fastest-growing skills included AI Agents and LangGraph · source
statfern. "AI at Work Statistics 2026." Published 2026-11-13. https://statfern.com/ai-work-statistics/
The primary sources behind this report. We link to the page where each figure appears.
| 01 | McKinsey & Company (QuantumBlack) | McKinsey Global Survey on the State of AI, N=1,993 respondents at all levels, Jun 25–Jul 29, 2025; firm-reported deployment measure (not individual employee usage) | View source → |
| 02 | Microsoft (2025 Work Trend Index) | 2025 Work Trend Index: survey of 31,000 workers and leaders across 31 countries plus LinkedIn labor-market data and Microsoft 365 telemetry; leader-reported | View source → |
| 03 | Gallup | Gallup Panel survey of 22,368 employed U.S. adults, Oct 30–Nov 14, 2025; self-reported individual usage (different 口径 from firm-level surveys) | View source → |
| 04 | U.S. Census Bureau (Business Trends and Outlook Survey) | BTOS biweekly nationally representative survey of ~1.2M businesses; AI question revised Nov 2025 to cover 'any business function', so not comparable with the narrower pre-2026 'producing goods or services' series; firm-level measure | View source → |
| 05 | Stanford HAI (AI Index Report 2026) | Synthesis of McKinsey Global Survey data (org-reported); Europe 91%, North America 90%; agent use measured separately and still early | View source → |
| 06 | Deloitte (The State of AI in the Enterprise) | Annual survey of 3,235 director-to-C-suite leaders across 6 industries and 24 countries, Aug–Sep 2025; access measure ≠ active usage | View source → |
| 07 | Brynjolfsson, Li & Raymond — Quarterly Journal of Economics (NBER WP 31161) | Field data from 5,179 customer support agents; productivity = issues resolved per hour; published QJE 140(2):889–942; effect smallest for experienced/high-skilled agents | View source → |
| 08 | Noy & Zhang — Science | Randomized online experiment, 453 college-educated professionals, occupation-specific incentivized writing tasks; quality rated +18%; inequality between workers decreased | View source → |
| 09 | Upwork Research Institute | Global survey of 2,500 workers (1,250 C-suite, 625 full-time employees, 625 freelancers), conducted Mar 25–Apr 9, 2025; self-reported, not measured output | View source → |
| 10 | Pew Research Center | Survey of 5,273 employed U.S. adults, Oct 7–13, 2024; self-reported helpfulness among chatbot users | View source → |
| 11 | Goldman Sachs Research | Model-based estimate extrapolating U.S./Europe occupational task-content data to employment-weighted global totals; 'exposed to automation' ≠ 'displaced' — most exposed jobs are only partially automatable | View source → |
| 12 | Stanford HAI (AI Index Report 2026, citing McKinsey 2025 survey) | Employer expectation measure; anticipated reductions highest in service operations, supply chain and software engineering; large-scale job losses not yet visible in aggregate employment data | View source → |
| 13 | LinkedIn Economic Graph (AI Labor Market Update) | LinkedIn hiring data across US/UK/France/Germany/India, last-three-months window; 1.3B+ member base; hiring-demand measure, not employment levels | View source → |
| 14 | World Economic Forum (Future of Jobs Report 2025) | Projection extrapolating a survey of 1,000+ employers (55 economies, 14M workers covered) with ILO global employment data; driven by macrotrends beyond AI (demographics, green transition, geoeconomics) | View source → |
| 15 | McKinsey & Company | The State of AI in 2025 survey, N=1,993, Jun 25–Jul 29, 2025; expectation measure — a small share also expected headcount to increase | View source → |
| 16 | Lightcast (data partner, Stanford AI Index 2026) | Analysis of billions of U.S. job postings; labor-demand measure; +72% vs 2022, +297% vs a decade ago | View source → |
| 17 | MIT NANDA (The GenAI Divide: State of AI in Business 2025) | Multi-method study: 300+ public AI implementations analyzed, 52 structured interviews, 153 senior-leader survey responses; returns measured as measurable P&L impact | View source → |
| 18 | Deloitte (State of Generative AI in the Enterprise, Q4 report) | Wave-four survey of 2,773 director-to-C-suite leaders across 14 countries, conducted Jul–Sep 2024; nearly three-quarters also said their most advanced GenAI initiative is meeting or exceeding ROI expectations | View source → |
| 19 | Upwork Research Institute (Future Workforce Index 2025) | Survey of 3,000 U.S. skilled knowledge workers, Dec 2024–Feb 2025; freelancers also use AI more often (62% several times a week vs 53% of FTEs) | View source → |
| 20 | LinkedIn Economic Graph (The Skills Signal Report 2025) | LinkedIn Economic Graph modeling; expansion factor for AI talent pipeline under skills-first hiring | View source → |
| 21 | Lightcast (Stanford AI Index 2026 data partner) | New Agentic AI skill cluster tracked in 2025 (Lightcast now tracks 300+ AI skills in 10 clusters); fastest-growing skills included AI Agents and LangGraph | View source → |
Data editor, statfern. Every statistic on this page was opened at its primary source before publication. Editorial policy →
Need numbers we haven't published?
Bespoke statistics reports for teams and publications. The same rule applies to every project: each number carries a primary source.