从试验到价值：生成式人工智能推动企业重塑

From experimentationto impact Authors Jack AzaguryGroup Chief Executive– ConsultingLinkedIn Mike MoorePrincipal Director– Accenture ResearchLinkedIn Lan GuanChief AI OfficerLinkedIn Muqsit AshrafGroup Chief Executive– StrategyLinkedIn Sen RamaniLead – Data & AI, GlobalLinkedIn Making ReinventionReal with Gen AI:From experimentationto impact But perhaps the biggest barrier concerns talent.Our research found that 3x more gen AI budgetsare spent on technology than on people.3Whiletechnology anchors any transformation, it’s thealignment of people, processes and technologythat drives reinvention. The five imperatives—one year on 3Build an AI-enabled, securedigital core 1Lead with value Definition:Focus on bold, high-impact initiativesand business cases that drive reinvention in coreprocesses. In this context, how can organizations overcomethese challenges—and make reinvention trulyreal—in the age of gen AI? This report reflectson the lessons we have learned from the pastyear of gen AI project delivery and providesactionable insights for leaders working toimplement and scale this epochal technology. Definition:Build the infrastructure needed to scalegen AI effectively while creating differentiation. Twelve months ago, inReinvention in the Age ofGenerative AI, we forecasted that 2024 wouldmark a shift from “education and experimentation”to building the foundation for generative AI(gen AI) and delivering measurable value. Wealso introduced five imperatives—key areas fororganizations to prioritize as they seek to harnessgen AI and create material impact (see call-out box). What’s new:Gen AI’s ability to “hack workflows,”tackle organizational silos and disrupt traditionalbusiness models and value chains can help tounlock growth from new services, channelsand experiences. By driving integrated, mega-process transformations, these capabilities helporganizations address core challenges in innovativeways and deliver greater end-to-end impact acrossprocesses. What’s new:Agentic architectures and modularAI platforms are unlocking new levels of scalable,always-on intelligence across the enterprise. 4Close the gap on responsible AI 13%of executives reportcreating significantenterprise-level value.Just Definition:Embed governance, ethical frameworksand monitoring to balance innovation withaccountability. Since then, we have analyzed over 2,000 gen AIprojects delivered to clients, conducted surveyswith more than 3,000 C-level executives anddrawn on insights from our own practitioners(see ‘About the research’ for further details). Thisresearch reaffirms the enduring importance ofthe five imperatives, showing that organizationsacting across all five are 2.5x more likely to realizeenterprise-level results.1 2Reinvent talent andways of working What’s new:With regulatory scrutiny of AIintensifying across the globe, the stakes are higherthan ever. Organizations must be “responsible bydesign” to build trust, mitigate risk and safeguardsustained innovation. Definition:Guide a vision for reinventing work,reshaping the workforce and preparing employeesfor a gen AI-powered world. 5Drive continuous reinvention Yet many still struggle—only 36% of surveyedexecutives say they have scaled gen AI solutions,and just 13% report creating significant enterprise-level value.2Challenges around data readiness,process redesign and a lack of C-level sponsorshipcontinue to hinder progress. What’s new:AI agents are transforming how AI andemployees collaborate, enabling organizations toembed AI as “digital colleagues” and reimagineworkflows, roles and capabilities. Definition:Build organizational agility to adaptquickly and stay ahead of disruption. What’s new:Rapid advancements in inferencingcapabilities, AI agents and physical AI are makingcontinuous change an enduring capability. This shiftunderscores gen AI's role as a critical enabler ofreinvention at speed across industries. Urgency to act The incentives to invest in scaling gen AI are clear.Leaders in IT, customer service and marketing, forexample, are deploying cross-functional solutionsthat streamline operations and elevate customerexperiences. Seventy-three percent of gen AIinvestments focus on these functional areas,with high-impact results in technology deliverylifecycles (TDLC), contact centers and marketingpersonalization.4 The urgency to act could not be clearer: Companiesalready delivering enterprise-level value are 4.5xmore likely to have invested strategically in agenticarchitecture and are 6x more likely to significantlyincrease their gen AI investments in 2025. Thesecompanies are positioning themselves to compoundtheir advantage; those still hesitating risk slippingfurther behind. In the sections that follow, we revisit the fiveimperatives—evaluating the progress organizationshave made in adopting them and exploring howthese lessons can help organizations acrossindustries overcome persistent challenges andachieve lasting value. Meanwhile, in industries