制造业中的人工智能

Systemic AI at theroot of manufacturingperformance The new growth infrastructurefor manufacturers Authors RolandMayr LuisLuque PrasadSatyavolu TraceyCountryman Supply Chain andEngineering GlobalLead, Accenture Industry andEnterprise, IndustrialsIndustry Group, GlobalLead, Accenture Manufacturing andOperations GlobalLead, Accenture OperationalTechnology SecurityGlobal Lead, Accenture Prasad Satyavolu shapesAccenture’s globalmanufacturing agenda,with a key focus ongrowing the firm’s PhysicalAI business across theAmericas, EMEA andAPAC. He brings morethan 30 years of executiveexperience in industry andprofessional services atthe intersection of cyber-physical systems and theconvergence of digitaland physical automation.He has led large-scaletransformations acrossmanufacturing, productdevelopment, supply chainand operations. His workis focused on advancingre-industrialization, frommajor CapEx programsand competitive products/platforms to Physical AI–enabled manufacturing. Tracey Countryman leadsAccenture’s global SupplyChain and Engineeringbusiness, shaping thefirm’s strategy, investmentsand innovation agendaacross industries. Shebrings a 30-year career inmanufacturing and supplychain transformation andhas helped build and scaleAccenture’s manufacturingand supply chaincapabilities over time. Herexperience spans networkstrategy, supply chainplanning, manufacturingtransformation andlarge-scale programsthat industrialize digitaland AI capabilities fromfoundational data andplatforms through plant-level execution. Luis Luque bringsmore than 25 yearsof experience at theintersection of industrialtechnology andcybersecurity. He leadsa cross-industry globalteam focused on securingcritical infrastructure, fromlegacy manufacturing andutilities environments to AIdata centers and roboticsinfrastructure, buildingrisk-based, AI-drivencapabilities for sensitiveindustrial settings. Priorto joining Accenture, heco-founded Cimation(acquired by Accenture),where he helped shapeindustrial IT, automationand ICS security services. Roland Mayr leadsAccenture’s globalIndustrials industry,advising executiveteams on reinventionwith digital, data and AIacross complex assetstructures and processchains. With over threedecades at Accenture, hebrings deep automotiveexperience and has heldsenior leadership rolesacross the firm, includingwithin Accenture Songand global client accountleadership. He is known fora strong customer focus,helping companies movebeyond products to deliverrelevant, meaningful andsustainable experiences. LinkedIn LinkedIn LinkedIn Page 7-8 Beyond the pilot, across the cycle Page 9-14 Defining systemic AI in manufacturing Page 15 The five key dimensions Page 16-17Dimension 1:Integrate planning, production,quality and logistics so decisions flow end to end Page 18-21Dimension 2:Build shared data, platforms andguardrails so you don't rebuild for every plant Page 22-23Dimension 3:Redesign decision rights andoperating rhythms so AI is part of daily work Page 24-27Dimension 4:Connect physical and agentic AIto create the closed loop Page 28-31Dimension 5:Design for humans in the leadto define accountability as autonomy scales Compounding advantages Page 32 The era of AI potentialis over. We’re now inthe era of AI scale. A global industrial technologycompany designed and testedone of its plants entirely in thevirtual world before laying a singlebrick. Concentrated, committedinvestment delivered decisiveresults: That company cut leadtimes by 78% and time-to-marketby 33%, raised productivity by 14%and reduced carbon emissions by28%. Net climate impact will varyby energy mix and the AI computefootprint over time. Across facilities operating today, audited results show that organizationsdeploying agentic AI in the supply chain recover from disruption 60% faster,carry 22% less inventory and reduce cost of goods sold by 5%. A leadingFrench automaker generates over $140 million per year from its AI program,while a Chinese appliance manufacturer estimated its AI program ROI overthe next several years at over 100%—with AI now embedded in 30% ofall employee KPIs. These results raise a practical question: How do manufacturers move fromisolated wins to an operating model that scales AI across a network? Toanswer it, we spoke with 36 manufacturing executives and experts acrossEurope, North America and Asia-Pacific. We wanted to understand howmanufacturers were embracing AI and rethinking how they did business withthese new technologies. What leaders at the top companies showed us wasfar more than a series of AI pilots: it was a whole new way of building theirorganizations. All stats and company examples are drawn from the interviewsunless otherwise cited. Here’s what we learned. Top performers have moved beyond fragmentedAI deployments to what we callsystemic AI: intelligence that operates asthe production system itself, not alongside it. It unites decision-making withphysical execution across machines, line