2025年全球人工智能调查报告

Content Excepted benefitsVS actual impact1 1Employeeenablement30 Risks & barriers17Conclusion34Get in touch A reality too good to be true? This year’s Global AI survey surfaces an image of AI adoption that seems almost too perfect. Ninety-nine percent of organizations report deploying AI; 70% place it at the heart of their business strategy;and IT budgets dedicated to AI have reached an average of 13%. Autonomous agents are alreadyproliferating and reported benefits span efficiency, customer value, and employee wellbeing. On paper,AI appears less like an emerging tool than a permanent fixture of corporate strategy. Yet many of theseachievements may be overstated. multiple agents requires new operatingmodels, accountability frameworks, andethical safeguards. Many are rushing toadopt, but few are prepared to govern. Behind the numbers: a fragile reality A market in transition Where the most advanced organizationsstand out by building resilient, industrial-grade foundations. They are investing indata architecture, governance layers, andadaptive platforms that transcend one-offexperiments. Others remain focused onquick wins, chasing new pilots and agentictools without rooting them in enterprise-wide processes. Sovereignty has becomea buzzword — cited by 84% — but in toomany cases it ends up assovereigntywashing, invoked rhetorically whileperformance, cost, or speed still dictatechoices. Beneath the headlines lies a morenuanced reality. Many initiatives still lackrigorous ROI measurement, projects oftenoverrun time and budget, and only about30% of users have meaningfully changedhow they work. Use cases are multiplyingbut most organizations are still onlybeginning to grasp how AI can trulytransform business models and coreprocesses. In many cases, thesedeployments echo what MIT calls the“GenAI Divide” — high adoption but lowtransformation, where visible pilots fail toscale. AI is now embedded in strategy, butmaturity remains uneven. Ambition isabundant; execution is inconsistent. In2025, the question will no longer be whoexperiments with AI, but who canprovevalue, scale responsibly, and anchor it intrust and purpose. Only then will AIbecome a true backbone ofcompetitiveness, not just another waveof corporate FOMO. This year’sWavestone AI Leader Survey,based on insights from 500 executivesacross Europe, North America, and Asia —CIOs, CDOs, CTOs, and CISOs — confirmsboth the centrality of AI and the paradoxof a technology celebrated in boardroomsyet fragile in execution. The real challenge ahead Imène KabouyaPartner,Wavestone At the same time, the rise of agenticorganizations - humans workingalongside intelligent agents - is redefiningexpectations from AI. This shift brings bothopportunity and peril: orchestrating About the survey Wavestone surveyed 500 mainlytechnology (but also cyber, data & AI)leaders based in the USA, the UK, France,Germany, Singapore, and Hong Kongin mid 2025. Respondent organizations’ size were:→23% large business(5000+ employees),→41% mid business(3000-4999 employees),→36% small business(1000-2499 employees). Consumer servicesBusiness & professionalservicesEnergy, oil/gas & utilitiesLife sciencesLeisure & entertainment IT, technology & telecomsRetail, distribution & transportFinancial servicesManufacturing & productionConstruction & propertyGovernment & internationalinstitutions 62% of respondents were Board/C-levelpositions and 38% were SeniorManagement and Director positions. Key takeaways AI maturity is not only about quick-wins, but about futureprooffoundations No strategy without AI Everyone wins with AI, on paper 70% of organizations already place AI at theheart of their business strategy While most organizations see tangiblebenefits, 46% do not yet have a structuredROI measurement framework. From everyday tools to infrastructure, leadingorganizations are laying the foundations forscalable AI 99% of respondents report that theirorganization has already deployed one ormore AI solutions, and 70% consider it asignificant part of their business strategy.With an average 13% of IT budgets dedicatedto AI, most companies are betting on short-term performance and long-termtransformation. The challenge ahead is toscale responsibly, with stronger governance,compliance and value tracking.> 99% of respondents report significant timesavings with AI, translating into businessvalue (32%), productivity (24%), andemployee wellbeing (24%). Its impact is setto be enterprise-wide, from core business tosupport functions, with the biggest gainsexpected in IT & cybersecurity (92%) andcustomer-facing functions (90%). ROI is notformally measured, but only 22% ofcompanies see it as a barrier to adopting AI.> AI adoption often starts with accessible,embedded tools (adopted by 89%), but themost committed organizations go further,laying the invisible foundations needed toscale. Investments in infrastructure andgovernance platforms enable them to movebeyond experimentation, buildinga re