《凱捷：2024生成式人工智能(AI)競品策略報告（英文版）（12頁）.pdf》由會員分享，可在線閱讀，更多相關《凱捷：2024生成式人工智能(AI)競品策略報告（英文版）（12頁）.pdf（12頁珍藏版）》請在三個皮匠報告上搜索。

1、Executive findings to inform strategy,governance,and AI investmentsGENERATIVE AI FOR A COMPETITIVE PRODUCT STRATEGYThis ebook examines how Generative AI has shifted the focus of business executives such as Chief Data Officers(CDOs)and Chief Data Analytics Officers(CDAOS),balancing opportunistic use

2、cases for building an AI-fueled competitive advantage with the challenges faced in responsibly operating and scaling AI.It combines findings from the Capgemini Research Institute with recent independent Domino Data Lab reports to evaluate how Generative AI has renewed the focus on building and opera

3、ting AI at scale.While expectations(especially from Generative AI)are high,particularly for product and service development,concerns over governance,security and responsible AI has most organizations fine-tuning open-source/commercial Generative AI models,with strategies looking towards fully in-hou

4、se large language models(LLMs).There are promising applications of Generative AI for product design and customer experience,in-house AI development brings massive challenges.Investment commitments fall short-particularly in the resources enterprise analytics leaders need to deliver on the promise of

5、 fine-tuned or in-house developed Generative AI models.Underinvestment in people,process,and technology is all-too-common.Data scientists therefore lack the data,toolsets and infrastructure necessary for them to do their job.This results in major talent retention and hiring challenges.More important

6、ly,the lack of access to proper tool sets increases the risk and exposure to governance and responsible AI issues.In conclusion,we suggest that organizations which lack urgency and commitment to effective people,processes,and supporting tools-like AI CoEs and AI platforms supporting hybrid-and multi

7、-cloud infrastructure-will likely lag behind market leaders.01020304055 Key FindingsGenerative AI has accelerated the trend of CDO/CDAOs shifting focus to business value creation,spurring new levels of investment.Governance,responsible AI,and security concerns have industry leaders building strategi

8、es for in-house LLM fine-tuning and development.While Generative AI brings massive competitive advantage to product design and customer experience,operationalization is difficult-and poorly governed models introduce massive risk.Critical shortages in centralized processes for collaboration and share

9、d best practices,lack of unified AI platforms(supporting AI infrastructure,data,and tools),and talent bottlenecks limit Generative AIs transformational potential.Companies that cannot build and operate their own AI models rapidly face existential competitive threats from those already integrating th

10、ese technologies into product and service offerings.EXECUTIVE SUMMARYGENERATIVE AI FOR A COMPETITIVE PRODUCT STRATEGY I Copyright 2023 Capgemini.All rights reserved.2CDOS/CDAOS TURN TO VALUE CRE ATION GENER ATIVE AI IS THE PATHCurrent level of integration of Generative AI into future product/service

11、 development plans Priority of driving new business results using AI/ML Top 3EmergingWe have already established a dedicated team and budget for its implementationWe are looking at establishing a dedicated team and budget for its implementation in this yearWe have not yet developed a concrete plan f

12、or integrationWe are currently unsure if or how we will integrate Generative AI into our product/service development plansNot importantTop priority30%46%16%8%3%8%49%40%of organizations say Generative AI is a topic of discussion in their boardrooms.of organizations have already established teams and

13、budget for Generative AI.76%of Data Science Executives(DSEs)“offense”=AI/MLMore than 3/4(76%)of data science executives(DSEs)see driving new business results with AI/ML as at least a TOP-3 priority for 2023.For nearly 1 in 3(30%)CDOs and CDAOs,this is#1 on their priority list for this year.98%40%GEN

14、ERATIVE AI FOR A COMPETITIVE PRODUCT STRATEGY I Copyright 2023 Capgemini.All rights reserved.3GOVERNANCE,RESPONSIBLE AI,&SECURIT Y ARE TOP CONCERNSDespite the Generative AI hype,governance poses a significant challenge58%of respondents(and 76%of C-level/VP data science executives)called governance o

15、ne of their most significant hurdles.Trust and responsibility must be built into AI SystemsIntellectual property,bias,inability to explain results,and inherited risk from underlying data remain top of mind.What are the biggest challenges to driving impact with Generative AI?Percentage of organizatio

16、ns that say the following are challenges for implementing Generative AI Lack of confidence that the Generative AI programs are fair(inclusive of all population goups)Responsible AICostInability to explain the results from Generative AI algorithmsControlDataBias in the Generative AI models leads to e

17、mbarrassing results when used by customers/clientsSkillsOperationalizationLack of clarity on underlying data used to train Generative AI programsGovernance58%51%51%45%48%49%37%44%49%36%41%GENERATIVE AI FOR A COMPETITIVE PRODUCT STRATEGY I Copyright 2023 Capgemini.All rights reserved.4AI Regulation i

18、s on the horizon,data regulation is hereThe E.U.AI Act,approved by European Parliament,in June 2023,will require Generative AI tool developers to submit their systems for review and publish copyrighted data used for model training before commercial release.,Data privacy regulations have also renewed

19、 focus on data residency(where is the data?)and data sovereignty(data is regulated by the laws of the country in which it is collected and processed).“Through 2026,nearly all multinational organizations will invest in local data processing infrastructure and services to mitigate against the risks as

20、sociated with data transfer.”As organizations invest in controlling data,they must also focus on controlling AI-by developing or fine-tuning open source or commercial LLMs themselves.BRINGING LMM FINE-TUNING AND DEVELOPMENT IN-HOUSEAI Teams Need Control for Responsible AI Balancing adoptions speed a

21、nd responsible AI obligations highlight the importance of customization and control in AI operations.In the near term,42%of respondents plan to fine-tune commercial and open-source Generative AI models.Longer-term,as they build in-house strategies and gain expertise,many plan to fully develop Genera

22、tive AI models from the ground up.94%39%Of AI teams believe they need to enhance off-the-shelf Generative AI offerings for their use casesOver one-third say they plan to steer their efforts fully towards in-house development of Generative AI.GENERATIVE AI FOR A COMPETITIVE PRODUCT STRATEGY I Copyrig

23、ht 2023 Capgemini.All rights reserved.5GENER ATIVE AI BENEFITS PRODUC T DESIGN&CUSTOMER EXPERIENCE.Generative AI benefits outweigh the risks74%of executives believe the benefits that Generative AI brings outweighs the associated risks.Anticipated benefits of Generative AI extend into product design

24、and customer experience.Fast movers are already capturing value Even before the recent Generative AI hype,Capgemini Research Institute analyzed the benefits gained by organizations with well-defined visions and strategies for intelligent products and services-data-fueled,smart,connected products(the

25、 precursor to Generative AI-powered products).of fast movers have seen a reduction in cost of servicing due to intelligent products&services.of fast movers have seen improved customer experience.of fast movers have accelerated R&D for improvements to existing products and services.83%83%79%Priority

26、of driving new business results using AI/ML Generative AI can enable us to create products and services that are more accessible and inclusive,serving a wider range of customers with diverse needs and preferencesGenerative AI can enable us to create more interactive and engaging experiences for our

27、customersGenerative AI can be used to improve customer service by providing automated and personalized supportGenerative AI can improve internal operations and enhance facility maintenanceGenerative AI will allow the design process to be more efficient and streamlined78%76%71%67%65%GENERATIVE AI FOR

28、 A COMPETITIVE PRODUCT STRATEGY I Copyright 2023 Capgemini.All rights reserved.6Insufficient technology and poorly governed data and AI models introduces risk,impeding product and service innovation.Big losses await poorly governed models CDOs and CDAOs have their eyes on many risks,but model govern

29、ance(i.e.,the models fueling product innovation),is a major money matter.Todays vast and quickly evolving regulatory landscape,paired with the high stakes of many enterprise data science initiatives,means that lack of trustworthy AI could cost companies tens of millions.EVEN THOUGH IN-HOUSE AI DEVEL

30、OPMENT BRINGS CHALLENGESof CDOs and CDAOs believe failure to properly govern their AI/ML applications would mean losses of$50 million or more to their companies.predicted losses of at least$10 million for their companies,and none predicted losses beneath$1m.44%87%Initiatives get stuck at pilot or Po

31、C stages Collaboration,skills,data,and technology-related challenges impede progress for intelligent products and services.Identified certain application areas/use casesImplemented pilots/proofs of concept of the identified use cases for at least one product/business linePartially scaled the identif

32、ied use cases for at least one product/service/business lineFully scaled the identified use cases for at least one product/service/business line7%13%48%31%GENERATIVE AI FOR A COMPETITIVE PRODUCT STRATEGY I Copyright 2023 Capgemini.All rights reserved.7.AND CRITIC AL SHORTAGES PL AGUE AI INNOVATIONLa

33、ck of CentralizationCollaboration,skills,data,and technology-related challenges put the brakes on progress.Lack of ProcessCDAOs look for unified MLOps platforms processes to provide automated governance capabilities,as well as the tools to improve AI productivity.Talent BottleneckTeams need to use t

34、he best tool for the job,not the one they are necessarily given:99%of data execs also agree that not providing talent with their preferred tools of choice negatively impacts the ability to hire,retain and/or upskill data science talent.%of respondents stating the following as challenges Compliance,s

35、ecurity&productivity also lead process needs Hiring&retention of data scientists is impacted by tool choice Absence of a single department/unit responsible for planning intelligent products and services for the entire organizationImproved data privacy and/or securityAbsence of an agile culture in Pr

36、oduction and Design teamsHigher levels of data science talent retentionImproved data science productivityMore and/or faster innovationMore direct business impact from data science supportDifferent organizational functions(including R&D,production,sales,etc.)working in silosGreater regulatory complia

37、nce30%30%46%46%16%16%8%8%8%45%Strongly Agree54%Somewhat Agree1%Somewhat DisagreeGENERATIVE AI FOR A COMPETITIVE PRODUCT STRATEGY I Copyright 2023 Capgemini.All rights reserved.8LOOKING AHE AD:KEEPING UP WITH COMPETITORS DELIVERYCompanies that cannot deliver products leveraging AI/ML or data science

38、models rapidly face existential competitive threat The promise of Generative AI launches executives into a sprint to make use of new,advanced algorithms upon their own data.98%of CDOs and CDAOs agree that the companies that bring AI and ML solutions to market fastest will be the ones to survive and

39、thrive in upcoming economic uncertainty.of Leaders use product usage and performance data to create new products/servicesof Leaders harness product usage and performance data to improve product qualityof Leaders use data from the field to carry out software updates and for future product iterations5

40、9%68%71%Leaders integrate data about product,usage,and customer behavior into ongoing product improvements GENERATIVE AI FOR A COMPETITIVE PRODUCT STRATEGY I Copyright 2023 Capgemini.All rights reserved.9ADDITIONAL INFORMATIONDive deeper into the research supporting this book Dominos CDO/CDAO Survey

41、 Report:Build a Winning AI Offense|C-level Strategies for an ML-fueled Revenue Engine.Go deeper into some of the statistics referenced in this ebook.Learn more about organizational challenges posed by IT owning data science platform decisions,the risk of poorly governed AI,and challenges presented b

42、y hybrid-/multi-cloud environments.Dominos REVelate Survey from June of 2023-a survey of AI professionals leading,developing,and operating AI initiatives across Fortune 500 companies.Capgemini Research Institutes Generative AI In Organizations report from early 2023 on top use cases across industrie

43、s.Capgemini Research Institutes Intelligent Products&Services report from April of 2022.Domino Data Lab provides a number of resources for data science leaders and practitioners,and for IT leaders who support and promote data science programs,including the following:Learn how Domino can help you sca

44、le Generative AI for the Enterprise:fast,safe,and economical.The Top 5 AI Considerations for Chief Data and Analytics Executives looking to accelerate enterprise data science in the hybrid cloud with MLOps The Domino Data Lab blog,featuring technical content,thought leadership,and strategic insights

45、 on the effective use of data science in all facets of business The Data Science Leaders podcast,featuring in-depth conversations with executives across industriesCitations1.DominoDataLab&WakefieldResearch,BuildaWinningAIOffense-C-LevelStrategiesforanMLFueledRevenueEngine,May20232.CapgeminiResearchI

46、nstitute,GenerativeAIinOrganizationsReport,April2023,N=800organizations3.Domino REVelate survey4.CapgeminiResearchInstitute,GenerativeAIinOrganizationsReport,April2023,N=800organizations5.TheWashingtonPost,“EuropemovesaheadonAIregulation,challengingtechgiantspower,”June14,20236.CNBC,“EUlawmakerspass

47、landmarkartificialintelligenceregulation,”June14,20237.Top5ConsiderationsforDataandAnalyticsExecutives,VentanaResearch8.Domino REVelate survey9.CapgeminiResearchInstitute,GenerativeAIinOrganizationsReport,April2023,N=800organizations10.CapgeminiResearchInstitute,Intelligentproductsandservicessurvey,

48、AprilMay2022;N=587organizationsthatalreadyhavewell-definedvisionsandstrategiesforintelligentproductsandservices11.DominoDataLab&WakefieldResearch,BuildaWinningAIOffense-C-LevelStrategiesforanMLFueledRevenueEngine,May202312.CapgeminiResearchInstitute,Intelligentproductsandservicessurvey,AprilMay2022,

49、N=1,000respondents from unique organizations that have or are currently building visions and strategies for a move to intelligent products and services13.CapgeminiResearchInstitute,Intelligentproductsandservicessurvey,AprilMay2022,N=1,000respondents from unique organizations that have or are current

50、ly building visions and strategies for a move to intelligent products and services14.DominoDataLab&WakefieldResearch,BuildaWinningAIOffense-C-LevelStrategiesforanMLFueledRevenueEngine,May202315.CapgeminiResearchInstitute,Intelligentproductsandservicessurvey,AprilMay2022,N=1,000respondents from uniqu

51、e organizations that have or are currently building visions and strategies foramovetointelligentproductsandservices,N=76organizationswhoareLeaders16.DominoDataLab&WakefieldResearch,BuildaWinningAIOffense-C-LevelStrategiesfor anMLFueledRevenueEngine,May2023GENERATIVE AI FOR A COMPETITIVE PRODUCT STRA

52、TEGY I Copyright 2023 Capgemini.All rights reserved.10ABOUT DOMINO&C APGEMINI Capgeminis AI,ML,&GenAI experts leverage Dominos enterprise-grade infrastructure for streamlined experimentation,model development,and production-grade coding,while Dominos integrated CI/CD pipeline amplifies development v

53、elocity.Moreover,Domino Data Labs enables the hub and spoke model to centralize development,reduce cost,and accelerate delivery with immediate ROI for data science teams,business units,and IT.Domino is the go-to centralized platform for collaboration,best practices,and innovation for any data scienc

54、e practice;including those that are starting to build and operate their own Generative AI models.”KOLIN KONJURAMLOps and GenAI Lead in AI&Analytics,Capgemini At Domino,we are immensely proud of our collaboration with Capgemini.Their profound expertise in AI,ML,and GenAI,combined with our robust Ente

55、rprise AI Platform,not only paves the way for innovative data science solutions but also empowers enterprises to realize tangible results and ROI.Together,weve cultivated an environment where cutting-edge technology meets seamless execution,ensuring that our collective clientele benefits from effici

56、ent,scalable,and trustworthy AI solutions.We are excited to push the boundaries of whats possible in AI and remain committed to setting the gold standard in model lifecycle management and time-to-value.SID KHAREGlobal Head of Partnerships,Domino Effective GenAI solutions are typically large,complex

57、models comprised of a cluster of smaller but similarly complex models.Putting GenAI models into production can scale very poorly without the proper infrastructure to deploy,and continuously monitor,retrain,and evaluate these solutions.Capgemini is building a strong,deeply-rooted partnership with Dom

58、ino Data Labs on which to deploy GenAI solutions for customers,giving them the visibility and control they need to generate ROI from their AI solutions.”PAUL INTREVADOGenAI Delivery Lead,Artificial Intelligence&Analytics,Capgemini North America Capgeminis collaboration with Domino represents a power

59、ful alliance for the development of enterprise-grade AI/ML solutions.By leveraging Dominos cutting-edge platform for model lifecycle management,Capgemini can streamline the development,deployment,and monitoring of AI models for our clients,all while allowing for effective cost management.This partne

60、rship allows us to harness the full potential of AI by ensuring model reproducibility,scalability,and effective governance,ultimately delivering more robust and sustainable solutions to our clients.Together,Capgemini and Domino Data Labs are at the forefront of driving innovation and value in the AI

61、 landscape.”AJAY MOHANGenAI Lead,Artificial Intelligence&Analytics,Capgemini North AmericaGENERATIVE AI FOR A COMPETITIVE PRODUCT STRATEGY I Copyright 2023 Capgemini.All rights reserved.11About DominoDomino Data Lab provides the Enterprise MLOps platform trusted by over 20%of the Fortune 100.Our pro

62、ducts enable thousands of data scientists to develop better medicines,grow more productive crops,adapt risk models to major economic shifts,build better cars,improve customer support,or simply recommend the best purchase to make at the right time.At Domino,our mission is to unleash the power of data

63、 science to address the worlds most important challenge.ContactKOLIN KONJURAMLOps and GenAI Lead in AI&Analytics,CSID KHAREGlobal Head of Partnerships,DPAUL INTREVADOGenAI Delivery Lead,Artificial Intelligence&Analytics,Capgemini North AAJAY MOHANGenAI Lead,Artificial Intelligence&Analytics,Capgemin

64、i North ACopyright 2023 Capgemini.All rights reserved.About CapgeminiCapgemini is a global leader in partnering with companies to transform and manage their business by harnessing the power of technology.The Group is guided every day by its purpose of unleashing human energy through technology for a

65、n inclusive and sustainable future.It is a responsible and diverse organization of nearly 350,000 team members in more than 50 countries.With its strong 55-year heritage and deep industry expertise,Capgemini is trusted by its clients to address the entire breadth of their business needs,from strategy and design to operations,fueled by the fast evolving and innovative world of cloud,data,AI,connectivity,software,digital engineering,and platforms.The Group reported in 2022 global revenues of 22 billion.Get the Future You Want|