0.8→1.7%of revenue now flows to AI — roughly doubling year over year, even as returns lag.
3%of organizations are scaling agentic AI across departments (900+ orgs, 15 industries, 10 countries).
88%organizational AI adoption in Stanford’s 2026 synthesis; its underlying surveys predate this September refresh.
~6%qualify as high performers: ≥5% EBIT impact and significant AI value.
67/32Organizational factors drive ~2× the AI impact of individual skill (67% vs 32%).
14%of enterprises have a documented AI strategy with clear goals. The rest default to cost-cutting.
77%of the hardest deployment challenges were invisible costs — change management, data quality, process redesign. Model choice was fully interchangeable for 42% of implementations, and a critical differentiator for only 19%.
7%run fully autonomous agents in production today (n=951).
80%+of CEOs are dissatisfied with progress on their AI transformation — from Bain's CEO survey, n=100, a much smaller base than the figures above.
15×year-over-year growth in active AI agents across Microsoft 365 — 18× at large enterprises (M365 telemetry, not survey).
74%of leaders expect to be using AI agents at least moderately by 2027 — few have scaled them yet.
58%of AI users are producing work they couldn't a year ago — 80% among Frontier Professionals.
40%of enterprise apps forecast to embed task-specific agents by end-2026. Published in 2025; not an observed adoption rate.
49%of Microsoft 365 Copilot conversations now support cognitive work — analysis, problem-solving, creative thinking (M365 telemetry, not survey).
21%of organizations have a mature governance model for agentic AI — agents are scaling faster than their guardrails.
19%of AI users work in the Frontier Zone, where individual skill and organizational readiness reinforce each other.
40%of large-company respondents report scaling agents in at least one function; 27% in 2025.
37%attribute positive enterprise EBIT impact to AI.
~20%report AI operating costs limiting use.
32%report forgoing a software purchase because coding agents enabled an internal build.
14 / 39%report past-year AI-related workforce declines / expect declines next year.
14%of surveyed CEOs define the profit-and-loss impact of every AI initiative.
26%of surveyed CEOs embed AI within broader business transformation.
59 / 9%have progressed beyond agent pilots / toward autonomous multistep workflows.
16%have widely or fully replaced fragmented legacy systems with an integrated platform.
20%have implemented AI testing, auditing, and risk-assessment processes.
74%of frontline white-collar employees use AI regularly.
41%of regular AI users report greater cognitive load.
42%of frontline regular AI users report saving at least one workday weekly.
66%of frontline regular AI users receive limited or no guidance on using saved time.
0.8→1.7%of revenue now flows to AI — roughly doubling year over year, even as returns lag.
3%of organizations are scaling agentic AI across departments (900+ orgs, 15 industries, 10 countries).
88%organizational AI adoption in Stanford’s 2026 synthesis; its underlying surveys predate this September refresh.
~6%qualify as high performers: ≥5% EBIT impact and significant AI value.
67/32Organizational factors drive ~2× the AI impact of individual skill (67% vs 32%).
14%of enterprises have a documented AI strategy with clear goals. The rest default to cost-cutting.
77%of the hardest deployment challenges were invisible costs — change management, data quality, process redesign. Model choice was fully interchangeable for 42% of implementations, and a critical differentiator for only 19%.
7%run fully autonomous agents in production today (n=951).
80%+of CEOs are dissatisfied with progress on their AI transformation — from Bain's CEO survey, n=100, a much smaller base than the figures above.
15×year-over-year growth in active AI agents across Microsoft 365 — 18× at large enterprises (M365 telemetry, not survey).
74%of leaders expect to be using AI agents at least moderately by 2027 — few have scaled them yet.
58%of AI users are producing work they couldn't a year ago — 80% among Frontier Professionals.
40%of enterprise apps forecast to embed task-specific agents by end-2026. Published in 2025; not an observed adoption rate.
49%of Microsoft 365 Copilot conversations now support cognitive work — analysis, problem-solving, creative thinking (M365 telemetry, not survey).
21%of organizations have a mature governance model for agentic AI — agents are scaling faster than their guardrails.
19%of AI users work in the Frontier Zone, where individual skill and organizational readiness reinforce each other.
40%of large-company respondents report scaling agents in at least one function; 27% in 2025.
37%attribute positive enterprise EBIT impact to AI.
~20%report AI operating costs limiting use.
32%report forgoing a software purchase because coding agents enabled an internal build.
14 / 39%report past-year AI-related workforce declines / expect declines next year.
14%of surveyed CEOs define the profit-and-loss impact of every AI initiative.
26%of surveyed CEOs embed AI within broader business transformation.
59 / 9%have progressed beyond agent pilots / toward autonomous multistep workflows.
16%have widely or fully replaced fragmented legacy systems with an integrated platform.
20%have implemented AI testing, auditing, and risk-assessment processes.
74%of frontline white-collar employees use AI regularly.
41%of regular AI users report greater cognitive load.
42%of frontline regular AI users report saving at least one workday weekly.
66%of frontline regular AI users receive limited or no guidance on using saved time.
