{"version":1,"asset_type":"statistical_series","data_type":"time_series","slug":"ai-enterprise-adoption","url":"https://apiardata.com/statistics/ai-enterprise-adoption","html_url":"https://apiardata.com/statistics/ai-enterprise-adoption/","title":"Enterprise AI Adoption — US, EU, and Global Business Use (2023–2025)","description":"Tracks the share of firms using artificial intelligence in business operations across three measurement frameworks: the US Census Bureau's Business Trends and Outlook Survey (BTOS), Eurostat's annual enterprise ICT survey (isoc_eb_ai), and McKinsey's State of AI global survey. Covers 2023 to 2025, the period of most rapid firm-level AI uptake.","domain":"papers","category":"AI & Research","keywords":["AI & Research","AI Adoption","Enterprise Technology","Digital Economy","Business Surveys"],"publisher":"US Census Bureau BTOS; Eurostat (isoc_eb_ai); McKinsey & Company State of AI Survey","frequency":"Annual","geography":"United States; European Union; Global (McKinsey survey)","temporal_coverage":"2023/..","last_updated":"2026-04-01","last_updated_text":"April 2026","data_as_of":"2025","variable_measured":"Share of enterprises currently using AI in production; share of EU enterprises using AI technologies; share of organizations using generative AI in at least one business function","measurement_technique":"Probability-sample business survey (US Census Bureau BTOS); stratified enterprise survey across EU-27 member states (Eurostat isoc_eb_ai); self-selected executive survey (McKinsey State of AI)","license":"https://apiardata.com/data-license","is_accessible_for_free":true,"sources":[{"name":"Federal Reserve — Monitoring AI Adoption in the US Economy (April 2026)","url":"https://www.federalreserve.gov/econres/notes/feds-notes/monitoring-ai-adoption-in-the-u-s-economy-20260403.html"},{"name":"US Census Bureau — Tracking Firm Use of AI in Real Time: Snapshot from BTOS (CES Working Paper 24-16R)","url":"https://www2.census.gov/library/working-papers/2024/adrm/ces/CES-WP-24-16R.pdf"},{"name":"US Census Bureau — AI Use Among Small Businesses (December 2024)","url":"https://www.census.gov/newsroom/blogs/research-matters/2024/12/ai-use-small-businesses.html"},{"name":"Eurostat — 20% of EU enterprises use AI technologies (December 2025)","url":"https://ec.europa.eu/eurostat/web/products-eurostat-news/w/ddn-20251211-2"},{"name":"Eurostat — isoc_eb_ai: Enterprises using artificial intelligence technologies","url":"https://ec.europa.eu/eurostat/databrowser/view/isoc_eb_ai/default/table?lang=en"},{"name":"McKinsey & Company — The State of AI 2025","url":"https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai"},{"name":"McKinsey & Company — The State of AI 2024","url":"https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-2024"}],"meta":[{"label":"Frequency","value":"Annual"},{"label":"Coverage","value":"2023–2025"},{"label":"Observations","value":"3"},{"label":"Geography","value":"United States; European Union; Global (McKinsey survey)"},{"label":"Last updated","value":"April 2026"}],"kpis":[{"label":"US Firms Using AI in Production (Census BTOS, Sept 2025)","value":"~10%","unit":"Share of all US businesses (narrow definition)","trend":{"direction":"up","value":"+6pp since late 2023"}},{"label":"EU Enterprises Using AI Technologies (Eurostat, 2025)","value":"20%","unit":"Share of EU enterprises with 10+ employees","trend":{"direction":"up","value":"+12pp since 2023"}},{"label":"Organizations Using Gen AI in At Least One Function (McKinsey, 2025)","value":"79%","unit":"Share of surveyed organizations (large-firm skew)","trend":{"direction":"up","value":"+46pp since 2023"}},{"label":"EU Large Enterprises Using AI (250+ employees, Eurostat 2025)","value":"55%","unit":"Share of large EU enterprises","trend":{"direction":"up","value":"vs. 17% for small firms (10–49 employees)"}}],"series":[{"key":"us_btos","label":"US (Census BTOS)"},{"key":"eu_eurostat","label":"EU (Eurostat)"},{"key":"global_mckinsey","label":"Global (McKinsey)"}],"observation_count":3,"observations":[{"period":"2023","us_btos":4,"eu_eurostat":8,"global_mckinsey":33},{"period":"2024","us_btos":6.5,"eu_eurostat":13.5,"global_mckinsey":71},{"period":"2025","us_btos":10,"eu_eurostat":20,"global_mckinsey":79}],"table":{"columns":[{"key":"segment","label":"Segment"},{"key":"us_btos_pct","label":"US BTOS (%)"},{"key":"eu_eurostat_pct","label":"EU Eurostat (%)"},{"key":"source","label":"Source"},{"key":"year","label":"Year"}],"rows":[{"segment":"Information (software & tech)","us_btos_pct":"22","eu_eurostat_pct":"","source":"US Census BTOS","year":"2025"},{"segment":"Professional / Scientific / Technical","us_btos_pct":"14","eu_eurostat_pct":"","source":"US Census BTOS","year":"2025"},{"segment":"Finance & Insurance","us_btos_pct":"12","eu_eurostat_pct":"","source":"US Census BTOS","year":"2025"},{"segment":"Manufacturing","us_btos_pct":"7","eu_eurostat_pct":"","source":"US Census BTOS","year":"2025"},{"segment":"Retail","us_btos_pct":"5","eu_eurostat_pct":"","source":"US Census BTOS","year":"2025"},{"segment":"Healthcare","us_btos_pct":"4","eu_eurostat_pct":"","source":"US Census BTOS","year":"2025"},{"segment":"Large enterprises (250+ employees)","us_btos_pct":"","eu_eurostat_pct":"55.0","source":"Eurostat isoc_eb_ai","year":"2025"},{"segment":"Medium enterprises (50–249)","us_btos_pct":"","eu_eurostat_pct":"30.4","source":"Eurostat isoc_eb_ai","year":"2025"},{"segment":"Small enterprises (10–49)","us_btos_pct":"","eu_eurostat_pct":"17.0","source":"Eurostat isoc_eb_ai","year":"2025"}]},"qa":[{"question":"Why does the US figure (10%) diverge so sharply from the McKinsey figure (79%)?","answer":"The two numbers measure fundamentally different things and should not be compared directly. The US Census Bureau BTOS asks businesses whether they are 'currently using AI in production' — a narrow operational test focused on deployed systems. It surveys all US businesses by size and industry. McKinsey's State of AI survey asks executives whether their organization uses generative AI in 'at least one business function' — a broader definition that includes piloting, non-core functions, and single-user deployments. Critically, McKinsey's respondent pool overrepresents large technology-forward enterprises: its 2025 sample consisted primarily of managers and C-suite executives at companies with revenues above $1 billion, and participation is self-selected, systematically attracting organizations that follow AI developments closely. The 69-percentage-point gap between the two figures reflects both the definitional breadth and the sampling bias of each survey instrument, not a contradiction in the data."},{"question":"What does the BTOS measure and what are its methodological advantages?","answer":"The Business Trends and Outlook Survey is a biweekly probability-sample survey operated by the US Census Bureau and Federal Reserve. Unlike most AI adoption surveys, BTOS uses a random probability sample from the Census Bureau's Business Register, which means its results are statistically representative of the full US business population by industry, firm size, and geography. Its narrow definition — 'currently using AI in production' — is also more economically meaningful than broader 'experimenting with' framings, because it captures deployments that have cleared an organization's implementation hurdle. The tradeoff is that the BTOS question covers any AI broadly (not just generative AI), and the narrow definition means the absolute share remains low even as uptake accelerates. Starting in November 2025 the Census Bureau revised the wording to a broader definition, which immediately pushed the reported figure to 17.3%; readings before and after that date are not directly comparable."},{"question":"What does the November 2025 BTOS definitional change mean for the trend series?","answer":"In November 2025, the Census Bureau broadened the BTOS question wording from a narrow 'currently using AI in production' framing to a broader definition that encompasses a wider range of AI-related activities. This caused the reported US adoption rate to jump from approximately 10% (the September 2025 reading under the old question) to 17.3% under the revised question. This is a survey methodology artefact, not evidence of a sudden doubling in real-world deployment. The series on this page uses only pre-November 2025 BTOS readings under the narrow definition to maintain a consistent time series from late 2023 through September 2025. Analysts building models that extend this US series forward will need to account for the break and use the November 2025 broader-definition figures as a new baseline."},{"question":"How does AI adoption vary by sector in the United States?","answer":"BTOS sectoral data for 2025 shows sharp stratification. The Information sector — which includes software firms and technology companies — leads at roughly 22%, nearly triple the all-business average. Professional, Scientific, and Technical Services follows at approximately 14%, reflecting heavy AI use in consulting, legal, and analytical services. Finance and Insurance firms report around 12%, consistent with early adoption of AI in credit, fraud detection, and client-facing automation. Manufacturing, Retail, and Healthcare trail at 7%, 5%, and 4% respectively. The sectoral gap illustrates that aggregate adoption figures mask a bifurcated market: technology-intensive sectors already have substantial AI integration, while the larger share of the business economy — small manufacturers, healthcare providers, and retailers — remains in early stages. This concentration also helps explain why the BTOS employment-weighted figures show higher rates in sectors with larger average firm sizes."},{"question":"How does EU enterprise AI adoption vary by firm size?","answer":"Eurostat's 2025 isoc_eb_ai data reveals a pronounced firm-size gradient that the EU aggregate figure of 20% does not fully convey. Large enterprises with 250 or more employees report a 55% adoption rate — more than triple the small enterprise rate. Medium enterprises (50–249 employees) sit at 30.4%, and small enterprises with 10 to 49 employees report 17.0%. The pattern reflects cost and capability barriers: large enterprises typically have dedicated data and IT teams, established vendor relationships, and the revenue base to justify AI investment. For deal teams assessing target companies, the implication is that firms with fewer than 50 employees in the EU are likely at the early stages of AI integration, while large enterprises in technology-intensive sectors may already be consolidating around a small number of AI platforms."},{"question":"What does the Eurostat isoc_eb_ai survey measure and how is it constructed?","answer":"The isoc_eb_ai series is drawn from Eurostat's annual Survey on ICT Usage and E-Commerce in Enterprises, conducted across all EU member states. It covers enterprises with 10 or more employees and asks whether the firm used any AI technologies in the reference year. 'AI technologies' in the Eurostat definition includes machine learning, natural language processing, speech recognition, computer vision, and robotic process automation — a broad but consistent definition applied annually across all member states. The 2023 figure of 8.0% and the 2025 figure of 20.0% are enterprise-count weighted, meaning each qualifying enterprise counts equally regardless of employment size; this differs from the BTOS which uses an employment-weighted framework. The Eurostat series is harmonised across EU-27 member states, making it the most directly comparable cross-country dataset for European AI adoption, though national implementations may vary slightly in how the survey is administered."}]}