《科爾尼：2024亞太地區電信業成功之路：把握生成式AI機遇實現轉型發展（英文版）（21頁）.pdf》由會員分享，可在線閱讀，更多相關《科爾尼：2024亞太地區電信業成功之路：把握生成式AI機遇實現轉型發展（英文版）（21頁）.pdf（21頁珍藏版）》請在三個皮匠報告上搜索。

1、Photo by Margo Stahl Kearney,New YorkScaling for success:how telcos are seizing GenAI opportunities in Asia PacificIn this paper,a collaboration between Kearney and Bridge Alliance,we shed light on how APAC telcos are grappling with these challenges,drawing from Kearneys Artificial Intelligence Asse

2、ssment(AIA)survey and from detailed discussions with more than 10 APAC telcosall Bridge Alliance member operators(BMOs).The paper focuses on what it will take to generate impact from AI with the telco organization.The stage is set for AI to uplift the way telcos are operating in APAC,helping to tran

3、sform how they use AI at scale and creating opportunities for immense economic value in the coming years.Telcos that swiftly implement the approaches outlined in this paper can optimize their market position and cement their leadership position within the dynamic AI ecosystem.As more companies acros

4、s a vast swathe of industries adopt AI,telcos around the world are certainly not resting idle.In fact,most have been early adopters of AI-powered models to support their businesses and have deployed predictive AI and machine learning initiatives such as churn prevention programs and“next best offer”

5、upselling for well over a decade now.The latest technologies advancements,such as machine learning,natural language processing,GenAI,and advanced data analytics,propelled by vast increase in GPU power performance over the past year and substantial investments in key enablers such as talent,storage,a

6、nd computing resources,are giving a renewed boost to such initiatives and fostering an environment that is ripe for much wider AI adoption.However,significant questions remain regarding the return on investment of certain initiatives and their scalability,particularly in the context of single-market

7、 deployments,and especially in smaller markets.Additionally,some Asia Pacific(APAC)telcos are facing unique challenges because of local market specificities,such as unique local languages and scripts,small market size,and lack of available talent.Foreword1Scaling for success:how telcos are seizing G

8、enAI opportunities in Asia PacificBased on our study,the telecom industry ranks seventh among 12 industries globally in overall maturity of AI adoption,behind industries such as consumer goods,technology,and retail(see figure 2 on page 3).This indicates that while the telecom industry is making stri

9、des towards integrating AI into its operations,there is still significant room for growth and innovation compared with other leading industries.Nevertheless,the die is cast,and some telcos around the world and in the APAC region are making big bets on AI,spurred by the push toward digital transforma

10、tion.AI is proving to be a game-changer,offering telcos the potential for top-line growth,improved customer experience,cost efficiencies and organizational effectiveness.Its applications span a wide range of areas,including customer service,network optimization,predictive maintenance,smart energy ma

11、nagement,and personalized targeted marketing to name a few.However,our research shows that there is also a wide disparity among telcos globally in terms of AI investment and maturity levels,with some preferring to adopt a wait-and-see approach before investing heavily into these new initiatives(see

12、figure 3 on page 4).Thinkbuildscale/govern:the three pillars of successfully scaling up AI executionIn 2024,Kearney conducted a global study of more than 1,000 companies across 12 key industries and more than 50 countries to gain insights into their data,analytics,and AI capabilities and maturity as

13、 well as how they compare to other players within or outside their industry.The study featured a survey with more than 40 questions addressing topics such as potential AI use cases within their organizations,the maturity of their current AI capabilities,and their planned investments in AI over the n

14、ext few years.Participant responses were rated on a scale of 04 based on the level of maturity.The aggregate of the responses was used to determine their maturity across the three building blocks of the AI lifecycle,based on Kearneys AI Readiness and Roadmap Framework:think,build,scale/govern.For th

15、e purpose of this paper,we have also supplemented Kearneys global study with an additional study of Bridge Alliance members and follow-on in-depth interviews with some of the BMOs CxOs.In Kearneys Artificial Intelligence Assessment 2024(AIA),only 4 percent of companies are classified as Leaders,whil

16、e 45 percent are Explorers,and 51 percent are Followers(see figure 1 on page 3).The key factors behind AI and analytics leadership include strong in-house talent capabilities,C-suite or C-suite-1 sponsorship,and early investment in data analytics and AI.The die is cast,and some telcos around the wor

17、ld and in the APAC region are making big bets on AI.2Scaling for success:how telcos are seizing GenAI opportunities in Asia Pacific3Scaling for success:how telcos are seizing GenAI opportunities in Asia PacificSource:Kearneys Artificial Intelligence Assessment Study 2024Figure 1Only 4 percent of com

18、panies fall into the Leaders category in Kearneys Artificial Intelligence Assessment Study 2024 Maturity distributionFollowers51%45%4%ExplorersLeadersFigure 2The telecom industry ranks seventh when it comes to maturity of AI adoptionOverall maturity of AI adoption across industries globally2.532.592

19、.672.692.692.702.712.722.752.752.792.832.85ChemicalsOthersEnergyInsuranceHealthcareAutomotiveTelecomMetals and miningRetailTransport and logisticsBankingTechnologyConsumergoodsSource:Kearneys Artificial Intelligence Assessment Study 2024Figure 3Three AI archetypes are emerging in the global telecom

20、landscapeKey characteristicsExample playersMaturity of AI and machine learning visionNote:Leaders ambitiously drive business transformations with bold AI and machine learning investments and pioneering new use cases.Explorers expand cautiously on existing AI and machine learning efforts with moderat

21、e risk,balancing innovation with value.Followers wait for AI adoption to mature,focusing on foundational improvements while managing legacy systems.Sources:company websites;Kearneys Artificial Intelligence Assessment Study 2024Followers51%Leaders(4%)Explorers(45%)Bold ambition for fundamentally chan

22、ging the business model Strong investments in AI,GenAI,and machine learning First movers in visionary use cases Commonalities:modernized data infrastructure and culture of innovation Telefonica(Europe and Latin America)Several advanced experiments,e.g.,Telefonica Aura,T Tech SK Telecom(South Korea)T

23、ransforming into a global AI leader with cutting-edge infrastructure and AI-powered services Spark(New Zealand)AI-powered decision model to maximize return on marketing spend AT&T(United States)Building on the strong 10+year AI and machine learning journey Taiwan Mobile(Taiwan)AI in telemarketing,cu

24、stomer service,and cybersecurity Telstra(Australia)AI and machine learning use cases in contact center and productivity Deutsche Telekom(Europe and United States)Co-investing to make custom LLM Globe Telcom(The Philippines)Isolated use cases focused on internal productivity,focus on letting employee

25、s experiment Vodafone Idea(India)One or two isolated use cases,current focus on business headwinds Smartfren Telecom(Indonesia)No notable AI and machine learning investments,focus on growth Cautiously experimenting with GenAI while continuing to be value focused from traditional AI and machine learn

26、ing Incremental builds on top of existing efforts in AI,machine learning,and advanced analytics Commonalities:large scale,moderate risk appetite,high complexity Waiting for the industry adoption to mature Establishing data foundations first Commonalities:risk-averse,dealing with legacy systems4Scali

27、ng for success:how telcos are seizing GenAI opportunities in Asia PacificTo support companies in their AI journey,Kearney has developed the AI Readiness and Roadmap Framework,broken down into three building blocks across the AI life cycle:think,build,scale/govern(see figure 4).This framework is desi

28、gned to help companies harness the full potential of their AI initiatives by ensuring that all AI investments and initiatives are designed and executed comprehensively across the organization.In the AI adoption journey,the think phase is the cool part,where companies define a bold AI vision and iden

29、tify impactful use cases that drive innovation.The build phase presents the hard challenge of aligning data and technology enablers to support this vision effectively.Ultimately,the scale/govern phase represents the hardest challenge,as organizations strive to embed AI deeply into their core operati

30、ons and culture for sustainable success.The AIA shows that APAC telcos exhibit lower overall AI adoption maturity compared with global and APAC telco leaders across all stages of AI maturity(see figure 5 on page 6).Yet,with business under renewed pressure from hyperscalers,telcos must act quickly an

31、d scale their key AI initiatives to establish themselves as frontrunners in the rapidly evolving AI landscape.The competition is intense,with other ecosystem players,including tech giants,already capitalizing on their strengths to offer comprehensive AI solutions on a global scale,which could curtai

32、l telco growth opportunities that AI offer beyond internal productivity programs and tools.Figure 4Kearneys AI Readiness and Roadmap Framework supports companies in their AI journeysThink:the cool part Build:the hard part Scale/govern:the hardest part Notes:B2C is business to consumer;B2B is busines

33、s to business;B2P is business to people;R&R is recognition and reward.Source:Kearney analysisVision:state-of-the-art use casesVision:Become the leading data-and GenAI-driven telco safely tapping the GenAI dividend as a trusted partner to customersAI-ready tech stack:technology to build,deliver,and r

34、un GenAIAI operating model:embedding GenAI within and across the organization and processesDrive top-line growthDrive the customer experienceDrive cost efficiency and organizational effectivenessAI use cases:tap measurable GenAI potential across the organizationB2CB2BB2PNetworkITHRFinanceLegalTech s

35、tack and toolsScalable architecture to enable multi use cases,and optimize cost and performance Data readinessData availability,accessibility,and compliancePartnership strategyStrategy to develop partnership ecosystemOrganization and operating modelOrganizational structure,ways of working,and R&RsPe

36、ople and cultureData and GenAI capability planning and talent strategyResponsible AI governanceNecessary policies,governance,audits,and risk mitigation processes in place5Scaling for success:how telcos are seizing GenAI opportunities in Asia PacificThink:telco AI initiatives are not one-size-fits-al

37、lBased on Kearneys AI Readiness and Roadmap Framework,use cases fall into three value-driver categories that impact each part of the telcos business model(see figure 6 on page 7):Top-line growth.AI initiatives that can help prevent revenue erosion(such as churn prevention)and fuel top-line growth,bo

38、th for their own existing telco services and for new customer-facing services,such as upselling their own telco services,“next best offer.”Additionally,some telcos are also using AI to fuel top-line growth via external service monetization opportunities,such as offering customer care chatbots for en

39、terprise customers and,on the AI infrastructure side,providing GPU-as-a-service(GPUaaS)to enterprise customers or a SK Telecom-led global telco AI alliance or AI-ready data centers for hyperscalers.Customer experience.AI initiatives that can help deliver superior customer across the board(such as af

40、ter-sales preventive services)Cost efficiency and organizational effectiveness.AI initiatives that can help deliver both capex and opex savings(such as a network roll-out or network management-based energy savings)Our research reveals that global AI leading telcos have been actively developing and d

41、eploying AI solutions across all three categories.However,most APAC telcos have been focused on customer experience and cost efficiency initiatives.For instance,Taiwan Mobiles recent investment in an IT services company aims to upgrade the telcos IT systems to support customer-facing applications to

42、 improve customer service and boost employee productivity.While some are starting to develop initiatives to fuel top-line growth,these are still mostly to support their existing telco services,with very few venturing into building new products and services for their business-to-consumer(B2C)and busi

43、ness-to-business(B2B)customers.Sources:expert interviews;Kearneys Artificial Intelligence Assessment Study 2024Figure 5Asia Pacific telcos are less mature in their AI adoptionAverage telco AI readiness scoreThinkScaleBuildGovern3.613.482.952.522.862.943.383.212.512.933.403.55APAC telco(all)APAC telc

44、o leadersGlobal telco leaders6Scaling for success:how telcos are seizing GenAI opportunities in Asia PacificFigure 6AI can create value for telco functions in a variety of ways 1 Cost reduction relative to overall telco opex2 Revenue uplift relative to overall telco revenueNote:Full employee rights

45、and responsibilities benefit displayed.B2C is business to consumer;B2B is business to business.Source:Kearney analysisCustomer-facingBackend operationsB2CB2B0.71%2revenue uplift0.10.3%1cost reductionAcquisitionPersonalization andomnichannel targetingProduct cross-sell/upsell RetentionNetworks1.72.8%

46、1cost reductionConfigurationmanagementCapacity managementFault and incident resolutionField operationsNetwork planning and investmentOperations and maintenanceAcquisition,whitespaceidentificationDeal pricingResource and salespipeline planning Retention1.72.4%2revenue uplift0.10.3%1cost reductionSupp

47、ort and corporateservices0.40.5%1cost reductionHR and workforce planningAdmin and overheadFinanceSupply chain and procurementCustomerservice0.81.3%1cost reductionTechnical support,includingcontact and call centersPost-sales customerengagementFront-line serviceIT,fraud,and security00.1%2revenue uplif

48、t0.30.4%1cost reductionFraudCybersecurityITThink7Scaling for success:how telcos are seizing GenAI opportunities in Asia PacificStudy results and discussions with APAC telcos reveal that they have adopted one of three main archetypes,in line with their AI maturity level and reflecting their appetite

49、for AI investment and the breadth of their AI strategies across their respective organizations:AI for the internal customerMost,if not all,telcos have already deployed and been using AI-based applications for many years,often with a focus on predictive analytics for churn prevention or“next best off

50、er”programs.All telcos in our study have been boosting these programs with new AI capabilities,which should be reflected in their top line moving forward.The telcos focusing on these applications only fall in the Followers category,preferring to adopt a wait-and-see approach:assessing whether ventur

51、ing and investing in AI initiatives farther away from their core will generate meaningful returns on their AI investments.AI for the internal and external customer within the existing telco operation footprintMany Bridge Alliance members are adopting this measured approach to AI initiatives with sel

52、ective investments focused on creating operational efficiency and an enhanced customer experience.Backed by its robust data and analytics capabilities,AIS has developed marketing models that they not only use for their own marketing campaigns,but also make available to some of their retail partners,

53、venturing into the field of marketing technology(MarTech).Similarly,Taiwan Mobile is offering its B2B customers its homegrown AI agent,which was initially developed for their own customer service center.Some of the more advanced players in this category,such as Singtel and AIS,are also making signif

54、icant investment in the infrastructure layer and have started to offer advanced enterprise services such as AI-ready data centers to hyperscalers and GPUaaS.In fact,Singtel has also expanded its GPUaaS offering to all the Bridge Alliance members,which can now in turn offer such service at an optimiz

55、ed cost to their own enterprise customers in their respective domestic markets.More recently,Singtel also launched RE:AI,an AI cloud service offering to democratize AI for enterprise to fuel innovation and growth in the region.Global AI strategy to reinvent the coreAt the most advanced end of the sp

56、ectrum,select telcos are placing AI at the heart of their overall corporate strategies as they aim to become AI companies.This cuts across most if not all functions,with investments in new AI-based product and services,not only to serve their own domestic market,but also to launch globally.These tel

57、cos are using AI as an opportunity to reinvent themselves and build a new core,leveraging the telco core capabilities and cash flow to fund new AI-based initiatives and ventures to support their transition from telcos to tech-cos.Here are two leading examples in this category:SK Telecom(SKT).The com

58、pany has positioned itself at the forefront of AI development not only in its domestic market and the APAC region,but also on the global scene.Last year,SKT announced the launch of its AI Pyramid strategy to become a global AI company,which relies on three pillars:AI infrastructure,AI transformation

59、(AIX),and AI service.Leveraging South Koreas deep engineering capabilities and talent pool and some of its sister companies in SK Group,SKT has taken the lead in developing some of its own AI infrastructure and AI-assisted services,which it intends to deploy globally,notably through the Global Telco

60、 AI Alliance,of which it is a founding member.China Telecom.With digital empowerment and AI as fundamental to the firms values,the company has advanced Chinas AI capabilities through the development of large language models(LLMs)with 1 trillion parameters that run on domestically manufactured chips.

61、Beyond the domestic market,China Telecom is looking to promote the overseas deployment of its mature capabilities in AI.8Scaling for success:how telcos are seizing GenAI opportunities in Asia PacificBuild:tough choices ahead to build the right foundations in a still evolving AI worldTelcos in APAC a

62、re spearheading with AI and GenAI use cases across a spectrum of opportunities.Whats interesting is how they are adopting different approaches to address the three fundamental build components of AI(see figure 7 on page 10):Tech stackWith the explosion of foundational models,even the AI tech stack h

63、as evolved.During our discussion and research with Bridge Alliance members,two important themes emerged for the AI tech stack,which has evolved with GenAI(see figure 8 on page 11):No one-model-to-rule-them-all.Almost all telcos see an AI tech stack that will include many foundational models across t

64、heir AI ecosystemboth small and large,text-based and multimodal,open-source,first-party,or even localized or built in-house.Smaller,one-shot models could be exceptionally useful for sentiment analysis,intent understanding,or other simpler tasks,while more complex models can have conversations with c

65、ustomers,employees,and more.For example,Globe is using both OpenAI and Gemini across a suite of GenAI use cases,and SKT has developed its own localized series foundational model,A.X,while both investing and collaborating with first-party providers such as Anthropic and OpenAI to secure a diverse mod

66、el lineup.Build-versus-buy,the age-old-question.Telcos are adopting two main approaches to GenAI infrastructure.Some,such as SKT,are building their own infrastructure,including GPU clusters and AI-centric data centers,to control their technology,prioritize data security and user privacy,and gain a c

67、ompetitive advantage in edge computing and low-latency AI applications.China Telecom has developed its own full-stack advanced cloud technologies to support and adapt to new AI use cases.Conversely,other companies,including Taiwan Mobile and Globe,are using existing cloud infrastructure from provide

68、rs such as GCP,AWS,Azure,and local players,including TWS for Taiwan.This strategy allows them to access powerful AI capabilities without the substantial investment required to build and maintain their own infrastructure,enabling them to prioritize AI application development for areas such as custome

69、r service and speech recognition.Telcos in APAC are spearheading with AI and GenAI use cases across a spectrum of opportunities.9Scaling for success:how telcos are seizing GenAI opportunities in Asia PacificFigure 7APAC telcos are adopting various approaches to address the three“build”components of

70、AINote:BMOs are Bridge Alliance member operators.Sources:expert interviews,secondary research;Kearney analysisTech stackData readinessPartnership strategyNo one-model-to-rule-them-allTech stacks dependent on firm needs GlobeLeverage OpenAI and Gemini SK TelecomDevelop own series foundational model w

71、hile collaborating with Anthropic(investment)and OpenAI Build-vs-buy,the age-old-questionBuild to maintain control or buy to minimize infrastructure investment SK TelecomBuild own infrastructure,including GPU clusters and AI-centric data centers Taiwan MobileLeverage providers(e.g.,GCP,AWS,Azure,TWS

72、),develop in-house if coreCentralized approach to AIDedicated team ensuring single-source-of-truth AISCentralized team that builds data foundations and ensures data governance of all GenAI solutionsFederated approach to AIIndividual teams work with data and analytics teams to build own AI solutions

73、GlobeDecentralized teams that build B2B AI agents and customer chatbots Hyperscaler partnerships For underlying infrastructure and support GlobePartner with Googles ready capabilities to develop AI applicationsEcosystem evangelismCollaboration with local stakeholders Taiwan MobilePartner with TWS(cl

74、oud provider),IT accelerator and research centersIndustry alliancesCollaboration among telcos SingtelPartner with Bridge Alliance to bring GPU-as-a-service to other BMOs SK TelecomPartner with worlds leading telcos to form the Global Telco AI Alliance(GTAA)Fundamental themes in the AI landscapeBuild

75、10Scaling for success:how telcos are seizing GenAI opportunities in Asia PacificFigure 8The AI technology stack has evolvedNotes:CRM is customer relationship management;RPA is robotic process automation;ITSM is IT service management;LLM is large language model.Source:Kearney analysisSimplified viewE

76、ngagementCore systemsInfrastructure and operations:compute,ITSM,etc.Enabling platformsCRM,RPA,other enabling platformsGenAI and LLMFoundational models,tuningtechniques,frameworksGenAI vendor landscapeData and analyticsMachine learning models andpredictive analyticsSelf-serviceWebsite,apps,3PAssisted

77、Shops,call center,field servicesIntegrationAPI ecosystemSecurityData extraction Chroma Pinecone Unstructured Weaviate ZillizOrchestration Haystack LangChain LlamaIndex Semantic KernelData extractions Arthur Guardrails AI Microsoft Presidio Rebuff TectonLLM providers Anthropic AWS Bedrock Cohere Goog

78、le Meta Open AIPrompt engineering HumanloopLLM ops Helicone MLflow Neptune.ai Replicate Weights&BiasesContent generation Amazon Polly Beatoven.ai Dall-e Midjourney Nova PhotoleapLLM cashe GPTCache RedisManaged implementations H20.ai Scale Stability AISynthetic data Gretel Hazy Mostly AI11Scaling for

79、 success:how telcos are seizing GenAI opportunities in Asia PacificData readinessAt the core of AI use cases resides data,and data readiness is table stakes for APAC telcos.Based on our discussions and research with APAC telcos,the AI Leaders have high maturity across the data landscape,such as data

80、 governance,data management,and more,and they believe it provides the right to win with AI(see figure 9).Telcos are investing as much as 20 percent of new investments into data initiatives to consolidate their data trove,which include information from interactions and long-term relationships with mi

81、llions of customers,demographics,location and mobility information,network traffic information,and a unique position to combine these domains of information to deliver unparalleled insights.Given the central role that data analytics plays across the various functions of the telco organization as wel

82、l as some of the necessary sensitivities around data privacy and governance,most telcos have a centralized data core driving AI readiness,with various degrees of centralization when it comes to developing use cases.AIS has adopted a centralized approach toward ensuring data readiness with a dedicate

83、d team building the data foundations,ensuring data governance with all GenAI solutions emanating from the data and analytics team.This approach ensures a single source of truth for company-wide datasets and rigid governance to ensure consumer data privacy protection for instance.Meanwhile,Globe Tele

84、com advocates for a democratized approach toward AI data readiness.The company has a centralized data and analytics(D&A)team,but individual business units are empowered to work with this centralized team to build their own AI solutions,such as B2B AI agents,customer chatbots,and more using shared pl

85、atforms and resources.This approach is intending to foster innovation,greater advocacy for AI and a culture of agile product delivery.Both approaches have their own merits and are best fit depending on the telcos own level of AI maturity.However,a centralized approach to ensure data readiness for AI

86、 tends to ensure greater scalability for telcos to fully and successfully realize their AI strategies.Figure 9Leaders in AI have high maturity levels across the data landscapeData and analytics technical capabilities proficiency overview by maturity levels Followers70%47%48%58%35%52%63%60%54%53%47%5

87、1%88%65%65%68%53%35%ExplorersLeadersData governance and securityData ingestion and integrationData management and qualityData platform and storageForward-looking capabilities(e.g.,AI,GenAI,and machine learning)ReportingSource:Kearney analysis12Scaling for success:how telcos are seizing GenAI opportu

88、nities in Asia PacificPartnership strategyIn this constantly evolving space,telcos are leveraging partnership strategies to bridge capability gaps such as infrastructure,AI capabilities,and other tools.Three archetypes of partnerships are emerging:Hyperscaler partnerships.Telcos are partnering with

89、hyperscalers to capitalize on their existing GenAI infrastructure and expertise.Globe Telecom and Taiwan Mobile leverage strong partnerships with Google to access ready AI capabilities and build solutions.This approach enables telcos to focus on developing AI applications,such as customer service ch

90、atbots and speech recognition systems,while relying on hyperscalers for the underlying infrastructure and support.Ecosystem evangelism.Some telcos are investing in building a strong local ecosystem,collaborating with various local stakeholders,including local providers,AI start-ups,research institut

91、ions,and government agencies.Taiwan Mobile exemplifies this model through its strong partnership with TWS,a local cloud provider,along with strong connections with a leading Taiwanese IT accelerator and its collaboration with local research centers to develop a speech recognition model specifically

92、for the Taiwanese market.This partnership approach allows telcos to tap into local talent,expertise,and data,enhancing their understanding of the local market and enabling the development of more relevant and impactful AI solutions.Industry alliances.Industry alliances are another way telcos are col

93、laborating with each other to provide early access to new technologies and capabilities to improve the AI ecosystem.Singtel is partnering with Bridge Alliance to bring its GPUaaS offerings to enterprises across the region,with other BMOs being early adopters gaining access to Singtels GPUaaS service

94、s.The Global Telco AI Alliance(GTAA),of which SKT is a founding party,is another strategic partnership between the some of the worlds leading telcos,with ambitious goals to harness the power of shared experience and expertise to deploy and sell telco-specific and geo-specific foundational models and

95、 AI solutions that will aim to outperform general-purpose counterparts.Scale/govern:seems to be the most challenging and often overlooked partScale/govern is crucial to unlocking AIs full potential while ensuring responsible deployment,but given their DNA,telcos still find scaling challenging.The mo

96、st common challenges include integration with legacy systems,data privacy concerns,and the ability to secure the right technical expertise and talent.To scale/govern AI well,telcos need to put in place the right organization and operating model,ensure robust programs around the people and change man

97、agement,and have a robust foundation for responsible AI governance.Organization and operating modelOne of the biggest challenges in scaling up AI is aligning the organization structure and operating model to support innovation and agility.Traditional hierarchies often slow decision-making,while insu

98、fficient collaboration between IT,data teams,and business units hinders AI adoption.13Scaling for success:how telcos are seizing GenAI opportunities in Asia PacificOur AIA study reveals that 43 percent of companies have adopted a hybrid AI operating model,37 percent a centralized model,and 20 percen

99、t a federated model.While companies may prefer one model over another,no single approach is universally superior.The choice depends on alignment with the organizations broader AI strategy(see figure 10).Companies that centralize innovation and decision-making often prefer the centralized model,while

100、 the federated model takes a more democratic approach by embedding AI into daily operations.However,our study also reveals that 100 percent of AI Leaders have C-suite or C-suite-1 leadership heading their AI initiatives,with 73 percent of their C-suite having backgrounds in analytics or AI.Each of t

101、hese operating models come with different benefits and drawbacks,based on our conversations with the Bridge Alliance members:Centralized model.Leaders with a global ecosystem-focused ambition tend to have a centralized team to rapidly build and scale a portfolio of high-value use cases.SKT has a ver

102、tically integrated AI team with dedicated engineers in both the United States and Korea,driving the organizations AI initiatives with a unified,cohesive push.By leveraging its central resources,SKT has been able to align its investmentssuch as AI-assisted services,pre-trained language models,and par

103、tnershipsacross the organization,ensuring efficiency,cost-effectiveness,and alignment with the companys overarching goals.Centralization also allows for scaling AI efforts,such as developing AI-centric data centers and launching the A.X language model,while maintaining control over data governance,p

104、rivacy,and compliance.Figure 10Choosing an AI operating model hinges largely on the companys broader AI strategyEmerging AI operating model archetypesScale/governNote:BU is business unit.Source:Kearney analysisCentralized(37%)Federated(20%)Hybrid(43%)Central AI teamGrowingAI teamCentral datagovernan

105、ce teamBUinitiativesownerCentral AI teamAI portfolioowners Centralized AI team evolves to become the core governance body AI portfolio owners(portfolio product managers)work with partners across business to validate high-value use cases Scaling responsibilities are pushed back into the business unit

106、s AI initiatives are driven by individual BU cases,leveraging the analytics and AI capabilities of the central data governance team Centralized AI team with representation from key units across the organization to rapidly build a portfolio of high-value use casesFunctionaland BUinitiativeowners14Sca

107、ling for success:how telcos are seizing GenAI opportunities in Asia PacificFederated model.In this model,expertise and governance are centralized,but scaling is decentralized across business units.Globe Telecom has adopted this democratized model,with a chief AI officer leading the AI enablement gro

108、up with a democratized approach guiding business units to scale AI use cases.We often see this model in telcos that are focused on using AI to boost internal productivity.Hybrid model.This model builds AI capabilities on top of an existing D&A team.At AIS,for example,the D&A governance team drives u

109、se cases,while a centralized D&A team provides support where needed by the business units.This flexible approach allows organizations to experiment as they scale AI while using their existing D&A infrastructure.Regardless of the operating model approach,the keys to successfully scaling AI alignment

110、are the business strategy,strong leadership,and frequent alignment with business users.People and cultureThe gap in capabilities in AI and data science talent is a common challenge globally,making AI adoption even more difficult.For example,SKT has proactively addressed this gap by evolving its in-h

111、ouse capabilities,positioning itself as a talent factory.Through focused investment in AI and data science expertise,SKT has not only advanced its domestic operations but also laid the groundwork for global expansion,using internal talent to fuel innovation and growth on an international scale.Talen

112、t availability has always been a struggle in adopting and scaling AI:60 percent of APAC telcos have identified the lack of technical expertise as a key barrier.This is compounded by the need to build LLMs for local languages.Globally,54 percent of analytics and AI Leaders are investing in specific r

113、ole-based training programs.Some examples of specific role-based AI training in the telecom sector include training for network engineers on AI-powered network management tools,customer service representatives on AI-driven virtual assistants,and marketing teams on predictive analytics for customer e

114、ngagement.Key training initiatives identified from discussions with BMO executives include the following:SKT,guided by its ecosystem strategy,emphasizes in-house talent development by targeting talent-rich markets with low costs per hire.This approach allows SKT to collaborate closely with leading e

115、xternal providers to enhance the companys AI capabilities.Taiwan Mobile noted that the CIO plays a crucial role in deciding whether to develop AI capabilities in-house or to outsource,including exploring mergers and acquisitions to bridge capability gaps.Building capabilities alone is not enough.Tel

116、cos also need the right culture to foster innovation and sustainably scale AI.Globe Telecom rolled out the AI Advocate initiative,where one in five employees is identified as an AI Advocate,receives AI training,and helps identify and drive AI use cases across the organization.In this way,AI adoption

117、 becomes an organic process fueled by business needs to adopt GenAI to enhance productivity.To accelerate scaling up AI,telcos can emulate Leaders by focusing on specific role-based AI training and working with external partners to train and acquire talent and ensure a fit-for-purpose culture to enc

118、ourage adoption.Responsible AI governance As trusted providers of national communication infrastructure,responsible AI governance is particularly important for telcos.As organizations seek to govern AI without stifling innovation,we offer a few considerations:Ensure strong data management practices

119、and introduce AI-specific governance and management tools.Complying with local laws and regulations on data privacy is of utmost importance;so is internal data management.Eighty-two percent leading organization have well-defined ownership of data quality,access,and compliance.Thirty-six percent of g

120、lobal telco Leaders have also introduced AI-specific governance,and 100 percent of Leaders have compliance training tailored for different functions.15Scaling for success:how telcos are seizing GenAI opportunities in Asia PacificTranslate responsible AI principles into practical steps.While many tel

121、cos have established guidelines for ethical AI use,these principles must be integrated throughout the organization.For example,data governance policies should specify which customer data is appropriate for specific use cases,ensuring privacy and compliance while minimizing ambiguity.By embedding suc

122、h practical guidelines,telcos can ensure that responsible AI is upheld at every level,from strategic decisions to day-to-day operations.Embed ethics at the heart of AI builds.Implementing clear responsible AI principles through trust and risk forums is essential for managing AI risks and ensuring re

123、gulatory compliance.Poorly managed AI can lead to significant financial,legal,and reputational damage,especially when data breaches or security lapses expose sensitive information.Additionally,AI models risk generating flawed or biased outcomesknown as“hallucinations”when trained on outdated or inac

124、curate data.This issue is worsened by the lack of model explainability,making it difficult to trace AI decisions and causing mistrust.To address these challenges,companies must establish robust AI governance and ethical guidelines early in the development process.In our study,we see two distinct app

125、roaches to funding AI investments.Some,such as SKT,follow a corporate venture capital funding model,using profits from core operations to strategically invest beyond its core systems to develop a telcoAI edge,for example,funding its acquisitions in Hynix and HBM.Through the Global Telco AI Alliance,

126、SKT collaborates with other telcos,sharing infrastructure costs and offering white-label AI-centric data center solutions.Conversely,more internal-focused telcos such as Globe Telecom follow a business-as-usual funding model,where AI projects are approved based on individual business cases,ensuring

127、alignment with strategic goals and delivering measurable outcomes.Bridge Alliance provides a compelling proposition to enable APAC telcos to scale for AI successBridge Alliance holds significant potential to help BMOs scale and govern their AI initiatives.By providing a unified AI readiness framewor

128、k and facilitating collaboration,the Alliance allows BMOs to adopt best practices and measure progress more effectively.Bridge Alliance already plays the role of value maximizer,facilitating sharing of AI use cases that offer the highest ROI potential.This is particularly of value to telcos with a s

129、ingle market focus,enabling them to focus AI investments on areas with proven business value given need for localized LLMs.Members can further benefit from sharing on how to align AI use cases,organization and operating model,and culture to their broader AI strategies.Bridge Alliance can further exp

130、and its role as platform and ecosystem developer,and as orchestrator.As a platform and ecosystem developer,Bridge Alliance can foster a collaborative environment that enables members to“build once,use multiple times.”This allows BMOs to efficiently share AI use cases,APIs,and infrastructure,helping

131、overcome talent shortages and resource constraints as well as allowing BMOs to implement AI solutions more quickly and at a reduced cost.The shared platform helps standardize efforts across BMOs,accelerating the AI maturity of less-experienced members and creating a democratized approach to AI devel

132、opment.Singtel has teamed up with Bridge Alliance to offer GPUaaS to enterprises across the region,with AIS,Maxis,and Telkomsel being early adopters of this services.Lastly,as an orchestrator,Bridge Alliance can play a vital role in forging partnerships among BMOs and global AI leaders,bringing cutt

133、ing-edge AI technology,products,and services to the region.By working with global Leaders,Bridge Alliance can help enhance the AI capabilities of BMOs,ensuring they remain competitive in the rapidly evolving AI landscape while accessing advanced solutions that help create technological advancements

134、and new revenue opportunities across APAC.16Scaling for success:how telcos are seizing GenAI opportunities in Asia PacificClosing noteThe massive increase in computing power performance over the past few months has certainly boosted AI and GenAI investments,fueling application and infrastructure dev

135、elopment in that space.Although it might still be too early to understand the full impact on the telco industryand other industries as well,for that mattertelcos are well-placed to play a pivotal role,both to boost their internal organizational effectiveness and to create new revenue streams beyond

136、their traditional telco services.In the results of Kearneys study and through our discussions with BMOs,we see some telcos taking the lead,carving out a place for themselves in the tech world as they transform from telcos to tech-cos.They are also using this opportunity to bring the industry togethe

137、r,sharing their AI infrastructure and their advancement in GenAI models and applications,supported by industry group such as Bridge Alliance,which offers a platform to share AI knowledge,use cases,talent,and infrastructure,ensuring that the lessons learned from the more mature and early adopters bri

138、ng the rest of the BMOs along on this AI journey.Telcos are well-placed to play a pivotal role,both to boost their organizational effectiveness and to create new revenue streams.17Scaling for success:how telcos are seizing GenAI opportunities in Asia PacificNikolai DobbersteinPartner,Kuala Lumpur Ol

139、ivier LetantPartner,Kuala Lumpur Achuth Pai Director,Innovation Projects Gareth PereiraPartner,Kuala Lumpur Ken WeeSVP,Partnership and New Business AuthorsBridge Alliance and Kearney would like to thank all the organizations that agreed to be interviewed for this paper.ReferencesChina Telecom https:

140、/www.chinatelecom- https:/www.chinatelecom- https:/ https:/ for success:how telcos are seizing GenAI opportunities in Asia PacificAbout KearneyKearney is a leading global management consulting firm.For nearly 100 years,we have been a trusted advisor to C-suites,government bodies,and nonprofit organi

141、zations.Our people make us who we are.Driven to be the difference between a big idea and making it happen,we work alongside our clients to regenerate their businesses to create a future that works for About Bridge AllianceBridge Alliance is the leading mobile alliance for premier operators and their

142、 customers in Asia Pacific,Europe,the Middle East,and Africa.The alliance today covers 35 member operators,which serve close to two billion customers collectively across these regions.Its goal is to build group capabilities and create value for members by enabling compelling roaming services and exp

143、erience,offering multi-market enterprise and IoT solutions,and delivering savings and benefits through leveraging group economies.Bridge Alliance covers these focus areas of work:5G,communications platform-as-a-service(CPaaS),gaming,generative AI starting with GPUaaS,IoT and enterprise mobility,grou

144、p sourcing,roaming,and telco APIs.Bridge Alliances members and partners include Airtel(India,Sri Lanka,and the Airtel subsidiaries in Africa:Chad,Congo,Democratic Republic of the Congo,Gabon,Kenya,Malawi,Madagascar,Niger,Nigeria,Rwanda,Seychelles,Tanzania,Uganda,and Zambia),AIS(Thailand),China Telec

145、om(Mainland China),China Unicom(Mainland China),CSL Mobile(Hong Kong),CTM(Macau),Deutsche Telekom(Europe),Globe Telecom(Philippines),Maxis(Malaysia),Metfone(Cambodia),MobiFone(Vietnam),Optus(Australia),Singtel(Singapore),SK Telecom(South Korea),stc(Saudi Arabia,Bahrain,and Kuwait),SoftBank Corp.(Jap

146、an),Taiwan Mobile(Taiwan),and Telkomsel(Indonesia).For more information,visit .For more information,permission to reprint or translate this work,and all other correspondence,please email .A.T.Kearney Korea LLC is a separate and independent legal entity operating under the Kearney name in Korea.A.T.Kearney operates in India as A.T.Kearney Limited(Branch Office),a branch office of A.T.Kearney Limited,a company organized under the laws of England and Wales.2024,A.T.Kearney,Inc.All rights reserved.