利用人工智能的价值：释放可扩展优势

Harnessing thevalue of AI Unlocking scalable advantage Table of contents 01 Who shouldread this reportand why? Business leaders across functions includingcorporate strategy, finance and risk, humanresources, marketing and sales, IT, sustainability,innovation/R&D, product design/development,sourcing, manufacturing/operations, and supplychain will find the report useful. It draws on thecomprehensive analysis of the findings from asurvey of 1,100 leaders (director level and above)at organizations with annual revenue above$1 billion across 15 countries: Australia, Brazil,Canada, France, Germany, India, Italy, Japan, theNetherlands, Norway, Singapore, Spain, Sweden,the UK, and the US. If you are looking to movebeyond experimentation to scaled, ethical, andhigh return on investment (ROI) AI deployment –this report is for you. This report offers critical insights for businessand technology leaders into the transformativepotential of traditional AI, generative AI (Gen AI),and Agentic AI. This applies to large organizationsacross sectors and industries includingautomotive, consumer products, retail, financialservices, telecom, energy and utilities, aerospaceand defense, high-tech, industrial manufacturing,pharma and healthcare, and the public sector/government. It is the third installment in ourannual research series and identifies shifts inadoption and investment levels, emerging usecases, as well as return-on-investment (ROI)trends. It also explores the implications of AIadoption for the workforce, governance bestpractices, and scaling strategies. AI is in transition. The technology is moving from isolated pilots tosweeping, mainstream adoption at an unprecedented pace. Organizationsare unlocking remarkable benefits, from accelerated innovation totransformative productivity gains. But this breakneck expansion also bringschallenges such as cost containment, workforce adaptation, and governancecomplexities, with a growing emphasis onethical, sustainable, trustworthy, and sovereign AI. This report encompasses the integration of technologies, includingtraditional AI, generative AI (Gen AI), and Agentic AI. Additionally, it analyzesthe implications for talent management and governance and suggestsstrategies to facilitate the responsible and sustainable scaling of Gen AI andAI agents. A new collaborativehuman-AI model The rise of AI agents Gen AI adoption is nowmainstream Investments in Gen AIare accelerating Already,14% of organizations haveimplemented AI agents at partial(12%) or full scale (2%), and nearlyone-quarter (23%) have launched pilots.Most (85%) executives are optimisticabout AI agents progressing to handleone or more business processes intheir functions over the next 3-5 years.Of the organizations already scalingAI agents, nearly 45% are piloting orscaling multi-agent systems, underliningthe pace of adoption. Almost four in tenorganizations today are optimistic thatAI agents will evolve into self-learningentities that require minimal humansupervision in the next 3-5 years. Gen AI adoption has surged from 6%in 2023 to 30% in 2025 -a fivefoldincrease in two years, with93%of organizations across sectors nowexploring, piloting, or partially/fullyenabling Gen AI capabilities. In telecom,for example, partially or fully scaledimplementation increased to 49% in2025, more than doubling from 2024.Customer operations, marketing, riskmanagement, and IT remain the leadingfunctions in Gen AI implementation.Gen AI adopters are also benefitingsignificantly, with around 80%expressing satisfaction with their GenAI outcomes. Nine in ten (88%) organizations arereportingan average 9% increasein Gen AI investments over the past12 months. Most (61%) executivesexpect this trend to continue over thenext year. However, cost concerns,including ‘bill shocks,’ remain prevalent.Organizations are also increasinglyturning to the cost-effectiveness ofsmall language models (SLMs). AI is evolvingfrom a tool into a teammember. Our research highlightsthat nearly 6 in 10 organizations arelooking at AI as either an augmenting/autonomous team member or asupervisor for other AI, with a viewto implementation in the next 12months – up from 44% currently. Butorganizations are underprepared forsuch a collaboration. Organizationsalso expect AI agents to focus onspecific tasks, rather than taking overentire roles, with almost 78% likely touse AI agents in this way in the next3-5 years. Nevertheless, more than 3in 5 organizations acknowledge thatemployees are concerned about theimpact of AI agents on jobs and careers. Generating value with AI: Architect the right processes, tech, data, and platforms for scalable AI: Trust around AI remainsan ominous concern AI’s growingenvironmental footprint –Redesign processes to deploy AI effectively–Develop a framework to identify the right blend of AI technology for theirbusiness and industry needs–Build robust data and tech ecosystems for scalable and responsible AI–Embrace ‘platformization’ fo