《電信管理論壇：2024通信服務提供商如何利用人工智能(AI)提升服務與創造價值：行業應用與戰略分析報告（英文版）（53頁）.pdf》由會員分享，可在線閱讀，更多相關《電信管理論壇：2024通信服務提供商如何利用人工智能(AI)提升服務與創造價值：行業應用與戰略分析報告（英文版）（53頁）.pdf（53頁珍藏版）》請在三個皮匠報告上搜索。

1、December 2024|www.tmforum.orgleveraging AIAuthor:Mark Newman,Chief AnalystEditor:Ian Kemp,Managing Editorsponsored by:for service andfor service andDecember 2024|www.inform.tmforum.orgSponsored by:Author:Mark Newman,Chief AnalystEditor:Ian Kemp,Managing Editor leveraging AIvalue creationTREND REPORT

2、Data&AI Innovationvalue creationfor service andDecember 2024|www.inform.tmforum.orgSponsored by:Author:Mark Newman,Chief AnalystEditor:Ian Kemp,Managing Editor leveraging AIvalue creationTREND REPORTData&AI Innovationcontents03 the big picture08 section 1:AI-enhanced network connectivity and service

3、s15 section 2:applications,services and products19 section 3:monetizing large language models 23 section 4:using platforms to deliver products and services 26 section 5:building an AI managed services strategy 31 section 6:key findings and recommendations35 additional feature:Cerillion unlocking AI-

4、powered telco product and service creation40 additional feature:Pegasystems the new imperative:transforming telecommunications for the AI era 44 additional feature:Red Hat paving the way to new revenue with an open and scalable AI platform 47 additional resourcesWe hope you enjoy the report and,most

5、 importantly,find ways to use the ideas,concepts and recommendations detailed within.You can send your feedback to the editorial team at TM Forum via editortmforum.orgbigpicturethe3inform.tmforum.orgThis research report focuses on the second category of value creation.However,it is not always possib

6、le to make a distinction between legacy and new.If AI adds value to an existing network in terms of the value that it delivers to consumer or business customers should this be seen as a legacy or a new service?For the purpose of this report we consider a range of AI-infused services,both new and enh

7、anced.We are interested,fundamentally,in the incremental revenues that communications service providers(CSPs)can generate from new or improved products and services.Categorizing new products and services opportunitiesIt is still too early to detect distinct patterns and trends in relation to the AI-

8、infused products and services that CSPs are building for their customers.Indeed,even the largest CSPs that we spoke to in the course of this research acknowledged that they were still in the relatively early stages of deciding where to put their efforts.However,we can glean enough from the announcem

9、ents they have made and their partnerships with technology vendors to build a picture of where they believe the best opportunities lie.Those opportunities are shown in the graphic opposite and summarized in the rest of this section.The benefits that AI can bring to telecoms operators fall broadly in

10、to two categories.First,there are gains that result in greater productivity and efficiency.These help operators to lower their capex and/or opex.Secondly,there are benefits that can be categorized as value creation.These result in higher revenues,either as a result of growing legacy services or,on t

11、he other hand,selling new services.4inform.tmforum.orgOperator AI product and service opportunitiesManaged Services(infrastructure,LLMs and applications delivered as services)PlatformsApplications,services and products(voice services,concierge services,security,IoT).Some developed in-house,some thir

12、d-party productsAI-enhanced network connectivity (work slicing,NaaS,CaaS)Foundational models /LLMs (e.g.Global Telco AI Alliance,Jiutian)AI infrastructure (compute,GPU-as-a-service,joint ventures with Nvidia)Telcos adding value through unique dataTM Forum,20245inform.tmforum.orgThe extent to which C

13、SPs are investing in building their own internal capabilities and in which part of the AI value chain is the main determinant of the products and services they are bringing to market.A number of CSPs globally are participating in the AI arms race by building data and cloud infrastructure for AI.Many

14、 are partnering with Nvidia to build so-called AI factories.However,others have decided not to build AI infrastructure.These tend to be the same operators that have adopted a public cloud first approach to cloud computing.Embedding AI solutions into the network,and into network operations,is somethi

15、ng that all operators will eventually introduce,either using their own internally-developed automation tools or by partnering with vendors or systems integrators.While the main objective of network automation,and the push towards autonomous networks,currently comes from efficiency gains,AI can also

16、help CSPs to develop more compelling networks and connectivity experiences.To monetize these experiences they will either need to charge higher prices,find new markets for connectivity in segments such as IoT,or bundle new products and services with connectivity.Very few operators have chosen to bui

17、ld foundational models for generative AI.The two most notable exceptions are Korean telco SK Telecom and China Mobile.SK Telecom spearheads the Global Telco AI Alliance with Deutsche Telekom,e&,Singtel and Softbank.Their mission is to build a multi-language telco-focused large language model(LLM)whi

18、ch they will market to other operators and service providers.China Mobiles Jiutian initiative is a little different.It is building LLMs for a number of different vertical segments within China.However,many more operators which have built platforms that sit on multiple LLMs and for their own internal

19、 GenAI deployments,are making these available to their customers.New operator opportunitiesApplications,services and products is a broad category that includes new AI-enabled applications that CSPs are launching as well as enhancements to existing services such as voice and text-based communications

20、.Some operators with strong in-house skills may try to innovate in this space,but a more likely scenario is that CSPs will resell third-party products and services in the same way that today they bundle video services or productivity tools with their communications services.We have identified enhanc

21、ed voice services,concierge services,IoT and security as some of the most interesting new opportunities for operators.When it comes to provisioning methods,platforms and managed services are two ways in which we see CSPs delivering different AI-infused services,particularly in B2B markets.CSPs inclu

22、ding Indian operator Jio,Singtel,Orange and Vodafone have built platforms for the delivery of a range of AI applications,products and services.But their approaches and the customer segments they are trying to target are all different to some degree or another.REPORTAuthors:Mark Newman,Chief Analyst,

23、TM ForumDawn Bushaus,Contributing Analyst,TM ForumSponsored by:Editor:Ian Kemp,Managing Editor,TM Forumestablishing links:platform models in the Open API economy March 2023Read our report to find out more about operators platform strategies:6inform.tmforum.orgManaged services,meanwhile,is a category

24、 in which most CSPs with large B2B businesses are already active.But even they may need to recruit new skills if they are to deliver AI managed services to their customers.As such,they may need to partner with AI specialists and hyperscale service providers,and choose the market segments where they

25、are most likely to be able to build a presence.Data is the differentiatorEven if CSPs can create new AI-infused products and services,and bring them to market,they will need to persuade their customers that they are trustworthy,value-for-money providers.The greater the differentiation in the product

26、s and services they provide the easier it will be for them to compete in the AI services market.But creating lasting differentiation in new products and services represents a major challenge for CSPs.The network itself is a differentiator,but more AI products and services will be delivered over the

27、top.Where CSPs can distinguish their products is in the data that feeds AI services.Any service built on data that is exclusive to the operator will offer differentiation.This could be data from the network itself,data built from the information that operators collect from customers that already sit

28、s in their IT systems,or the data that is captured from voice and text communications over operator networks.Even this may not be enough in itself,but if this data can be combined with and integrated into for example,the systems and services provided by hyperscale service providers,then operators st

29、and a good chance of realizing the opportunity.What plans do you have to use GenAI for building and enhancing products and services?60%50%40%30%20%10%0%New network services(work slicing/NaaS/CaaS)Sovereign infrastructure(e.g.GPU-as-a-service)Edge computing&resilienceAI solutions for B2B(integration

30、services,professional services,solutions marketplaces)Security/anti-fraud servicesEnhanced data protection&encryption(e.g.quantum networking)Voice value-added servicesn We are actively pursuing this opportunity n We are looking into this n We are not exploring this opportunityTM Forum,20247inform.tm

31、forum.orgB2B services are generating most CSP interestIn the summer of 2024 TM Forum conducted a global survey in partnership with Amazon Web Services(AWS)to gauge operators generative AI maturity.We identified a number of different product categories and asked AI decision-makers in CSPs to say whet

32、her they were actively pursuing these opportunities,looking into them or not exploring at the moment.Based on data from 203 AI decision-makers at 124 operators in 61 countries,the product lines that most CSPs said they are pursuing are voice services,and AI solutions for businesses such as professio

33、nal and integration services(see chart on the previous page).Read this report to:find out about the products,services and AI-infused solutions that CSPs are launching,or planning to launch,across their consumer and enterprise lines of business understand how CSPs can use AI to add value to network c

34、onnectivity explore the different partnership opportunities that are available to CSPs as they strive to innovate with new AI-enabled services learn where CSPS have unique capabilities that they can use to create sustainable value and differentiation assess which new revenue growth strategies are mo

35、st relevant and achievable for your business.The product lines that most CSPs said they are pursuing are voice services and AI solutions such as professional and integration services.AI-enhanced network connectivity and servicessection 18inform.tmforum.orgIn this section we consider the impact that

36、AI can make on the connectivity experiences that CSPs deliver to their customers and the opportunities for them to secure incremental and new revenues either by selling more,higher-priced,or services on top of,connectivity.When CSPs consider how and where AI can benefit the network and network opera

37、tions,generating new revenues ranks some way behind delivering efficiency and productivity gains and a better customer experience.That is illustrated in the graphic opposite,from our Benchmark report Building an AI strategy:telcos put the foundations in place,which included a survey based on respons

38、es from 104 executives from 73 CSPs globally.Indeed,the use cases that CSPs have deployed in the network in recent years have all been aimed at driving greater efficiencies.In 2024,any discussion about the use of AI in the network is inextricably linked with the drive to greater automation and,ultim

39、ately,the creation of autonomous networks.Reducing the number of people required to manage and operate networks is the key business driver.Whereas generative AI(GenAI)is emerging as the most promising and useful type of AI in many parts of the CSP business,it has tended to figure less prominently in

40、 discussions about network operations because of issues relating to accuracy and latency.Nevertheless,there are emerging use cases for AI in the network which will help to deliver greater value to network services.9inform.tmforum.org13%85%87%11%46%43%1%14%Where does AI/machine learning have the most

41、potential to positively impact the telecoms business?High potentialMedium potentialLow potentialNew product developmentOperations/network operationsCustomer experience/customer relationship managementTM Forum,2023Where does AI/machine learning have the most potential to positively impact the telecom

42、s business?TM Forum,202310inform.tmforum.orgWe have identified three categories of these use cases:Leveraging AI to help CSPs build new experiences and capabilities in the network services they offer their customers and for which they can charge higher prices or create new demand for connectivityUsi

43、ng AI to help CSPs better utilize their network capacity and,at the same time,free up spare AI computing capacity to sell to third partiesBuilding new network(and compute)capabilities based on enterprises requirements for originating and terminating AI traffic and monetizing this demand by selling n

44、ew capacity and SLA-driven experiences.There is,of course,huge overlap between these opportunities and trends.The new experiences and capabilities in network services that CSPs can offer their customers will help enterprises to better capture the opportunity to use AI in their businesses to deliver

45、new and enhanced experiences to their customers.But it is useful to consider them separately to better understand the full spectrum of opportunities that are available to CSPs.Delivering new network experiencesThe two key initiatives for CSPs to deliver new network experiences sit under the categori

46、es of network-as-a-service(NaaS)and network slicing.NaaS is a new consumption model for enterprises that allows them to operate their(private)networks and achieve the outcomes they need from them without owning,building or maintaining their own infrastructure.Those CSPs that wish to provide NaaS mus

47、t move their services and capabilities into the cloud and embrace new end-user pricing strategies and approaches that mirror infrastructure-,platform-and software-as-a-service.Despite the enthusiasm of many CSPs for NaaS,enterprise adoption has been extremely slow because of uncertainty around cost

48、and return on investment,and concern about how it will impact the day-to-day management of networks.However,analysts at ABI Research forecast that by 2030 more than 90%of enterprises globally will consume at least 25%of their network services via this model.Network slicing is a business model and a

49、network capability that was developed specifically within the context of 5G.The idea behind network slicing is that business users and different consumers have specific network requirements relating to location,time,speed,latency and throughput,and service-level-agreements more generally,and that un

50、til now mobile operators have been unable to meet these requirements.By migrating core network elements into the cloud,CSPs have the ability to meet these different requirements and,in doing so,deliver better experiences and,potentially,new revenues.But network slicing has been slow to build momentu

51、m.This is in part because CSPs have been slow to deploy new 5G core networks commonly known as 5G standalone(5G SA)that allow them to deploy network slices.The two key initiatives for CSPs to deliver new network experiences are network-as-a-service and network slicing.11inform.tmforum.orgAccording t

52、o Ericssons November 2024 Mobility Report,of the 320 CSPs globally with commercial 5G services only 60 have launched 5G SA.But slow network slicing progress is also due to the fact that it is not clear that CSPs public 5G networks can offer enterprises the SLAs they need.The question,then,is whether

53、 CSPs can use AI to improve their NaaS and network slicing offers and address some of the issues that are hindering end-user adoption.In the case of NaaS,AI and more specifically AI chatbots can play an important role in helping enterprises to monitor and manage their networks.“In the world of netwo

54、rk troubleshooting,AI-powered chatbots are a valuable tool that can provide immediate and around-the-clock support,”says Jamie McGregor,consulting manager at Deloitte in an article written for the website Medium.“Thanks to their ability to automate routine tasks and offer insights,AI chatbots can he

55、lp streamline network management and optimisation.They play a key role in diagnosing and resolving network and cyber issues,reducing downtime,and improving overall network performance.”While the use of AI was not initially baked into the network slicing concept it is now emerging as a tool that is u

56、seful both to CSPs and their customers to ensure the best possible match between the users requirements and what the slice delivers.The role of AI here is as an orchestrator,constantly analyzing and adjusting network parameters relating to factors such as latency,bandwidth and reliability to meet th

57、e specific requirements of the application.Dual-purpose AI and 5G networksThe success of Nvidia,the worlds largest provider of AI hardware and software,reflects the extent that CSPs are turning to AI in their network operations.Nvidia reported a 94%year-on-year increase in revenues to$35.1 billion f

58、or the third quarter to the end of October 2024 and a 110%year-on-year increase in operating profit to almost$21.9 billion.In the same period two years ago Nvidias revenues were$5.9 billion and its operating profit just$601 million.($millions)Q3 FY25 Q2 FY25 Q3 FY24 Q/Q Y/YRevenue$35,082$30,040$18,1

59、20 +17%+94%Gross margin 74.6%75.1%74.0%-0.5 pts +0.6ptsOperating expenses$4,287$3,932$2,983 +9%+44%Operating income$21,869$18,642$10,417 +17%+110%Net income$19,309$16,599$9,243 +16%+109%Nvidia financial results(GAAP)12inform.tmforum.orgNvidia has partnered with a number of companies including Nokia,

60、T-Mobile,Microsoft and Softbank in an initiative,called the AI-RAN Alliance,to develop dual-purpose AI and 5G networks.The company is also partnering with operators to build AI data centers and create a GPU-as-a-service line of business.In its third-quarter results Nvidia reported record quarterly d

61、ata center revenues of$30.8 billion,up 17%from Q2 and 112%compared to a year earlier.AI-RAN takes CSPs strategies for competing in the AI infrastructure arms race a stage further by embedding Nvidia compute technology in radio access networks(RANs).The basic premise for a combined AI-5G network is t

62、hat traditional RANs only operate at about one-third capacity,as they need to be over-provisioned to handle peak traffic loads as and when required.With a dual-purpose computing capability provided by AI-RAN,Nvidia believes that telcos now have the opportunity to monetize the remaining two-thirds ca

63、pacity forAI inference services and applications(see box opposite).It would be up to each operator to target different enterprises and market segments with its AI compute services.Japanese technology and telecoms group Softbank,for example,has set up a trial to use Nvidia AI technology to build real

64、-world AI inference applications,including autonomous vehicle remote support,robotics control and multimodal retrieval-automated generation at the edge.AI inference is the process of using a trained AI model to make predictions or decisions based on new,unseen data.In such instances the AI applies w

65、hat it has learned during training to real-world situations.This phase determines the effectiveness of AI in actual applications,ranging from recognizing speech to identifying objects in images.The inference phase comes after training,where the model learns from data sets by adjusting its parameters

66、.AI inference can be used in applications such as predictive analytics,computer vision to understand the content of images and videos,LLMs to understand and generate human-like text,and fraud detection to analyze transactions.Implementing AI inference typically requires powerful hardware such as CPU

67、s,GPUs(graphics processing units)and application-specific integrated circuits(ASICs)custom-designed chips optimized for specific AI models and tasks.They help variously with the speed,flexibility,performance and efficiency of AI inference implementation,but there are still challenges around latency,

68、accuracy and scalability.There are also different types of inference depending on application needs.Batch inference processes large datasets simultaneously,making it suitable for applications where real-time predictions are not needed.This approach is often used in periodic report generation and bul

69、k email personalization.Online or real-time inference processes data as it arrives,providing immediate predictions or decisions.This type of inference is crucial in applications such as autonomous driving,where sensors continuously feed data to the AI system,or financial trading to enable systems to

70、 react to market changes in real time.And streaming inference handles continuous,high-velocity data streams,such as those generated by IoT devices,social media platforms or live video feeds,or in applications such as smart city monitoring.Source:Run:aiWhat Is AI inference?13inform.tmforum.orgFor the

71、 next two to three years enterprises rather than consumers will be the main drivers of AI traffic growth.How AI is changing the face of telecoms networksIf the impact of AI on the digital lives of businesses and consumers is as big as the tech community expects,it will have a profound impact on the

72、size,shape and direction of telecoms network traffic flows.But there are many questions and issues that we need to work through before we can consider whether this traffic growth,and changes in traffic flows,represents a monetizable opportunity.They include:What will be the main applications that dr

73、ive AI traffic growth?Will the AI traffic mainly be confined to businesses rather than consumer services and applications?Where will AI traffic originate and terminate?What types of AI will generate what types of traffic?Will AI bring the edge computing opportunity to the fore and do CSPs have a rol

74、e to play?Does the growth in AI traffic help CSPs in the transition to NaaS and a gradual shift away from flat-rate monthly subscription pricing?For the next two to three years enterprises rather than consumers will be the main drivers of AI traffic growth as businesses consume new AI-infused produc

75、ts and services.These applications will primarily be used to automate low-value,repetitive tasks.B2B IoT services will also drive traffic growth in the emerging“sight as a sensor”category(see section 2).This B2B AI traffic growth will mainly be uplink traffic.But consumer AI traffic growth will inev

76、itably follow as entertainment and collaboration applications use AI to deliver new experiences.AI traffic originates and terminates in one,or more,of four locations:On a device such as a handset or a sensor On a premises,a company site and within a local cloud or data center At the edge,in a shared

77、 location In a regional or centralized cloud or data center.CSPs have the opportunity to generate new revenues if this traffic a)is generated from new devices such as IoT sensors,b)creates a demand for new or expanded fixed and mobile private networks or c)creates a demand for new capacity to connec

78、t businesses either to an edge or a regional or centralized data center.But there may also be an opportunity to generate revenues from the SLAs that AI users(businesses)will need to guarantee the performance and the security of their AI traffic.The(re)emergence of the edge opportunity is a function

79、of the need for some of the benefits offered by public cloud services in areas such as load balancing and high-volume traffic management,but combined with the low latency normally achieved with on-premises or public cloud computing.Many CSPs are now dusting off their edge computing business plans an

80、d repositioning them in the AI traffic era.Verizon,for example,has struggled to generate interest in its mobile edge services.But within the past six to nine months CEO Hans Vestberg has been talking up the prospects for the emergence of a GenAI applications edge computing market.In the next section

81、 we look at how CSPs can begin to create applications,products and services using AI.Applications,services and productssection 214inform.tmforum.orgTelecoms operators do not have a great track record when it comes to building new products and value propositions for their customers.Despite their very

82、 best efforts,most CSPs are still predominantly or completely dependent upon revenues from connectivity and communications services.This is not to say that CSPs do not offer other services to customers.But these are mainly products and services developed and owned by other companies for example,vide

83、o platform services(Netflix,Disney Plus and so on);productivity,collaboration or security tools(such as Microsoft 365 and unified communications applications);or hardware products(mobile phones,routers etc.).The main business rationale for selling these products and services is to help secure,or pro

84、tect,revenues from CSPs core products rather than to generate new revenue streams.Nevertheless,many CSPs are stepping up their efforts to bundle more third-party products with their own connectivity and communications services portfolio either through improved digital commerce capabilities or by cre

85、ating their own marketplaces.AI can help in two ways.First,it can help CSPs to generate greater insights into their customer needs.Secondly,the coming years will see an explosion in new AI-enabled products and services for CSPs to resell.In addition to selling services such as Microsoft Copilot to t

86、heir customers,CSPs have the opportunity to partner with lesser-known companies and to use their market reach and trusted customer relationships to find a much larger market than start-up companies could achieve themselves.15inform.tmforum.orgWe have identified a handful of product categories across

87、 B2B and B2C markets where CSPs may be able to provide new or enhanced services,either through their own innovation or working with partners.Enhanced voice servicesCSPs voice service revenues have collapsed over the past 20 years as customers have migrated to IP services and applications.As such,the

88、 idea that operators might be interested in investing in their voice business does,at first sight,seem unlikely.However,the argument could also be made that CSPs need to do something to protect the voice business and revenues they still have,otherwise the drift will continue.Operators voice services

89、 and text-based communications services have remained unchanged for many years and have been overtaken in terms of functionality by VoIP offered through services such as WhatsApp for consumers and Microsoft Teams in the The coming years will see an explosion in new AI-enabled products and services f

90、or CSPs to resell.16inform.tmforum.orgB2B market.The functionalities that these VoIP providers have added to their services include video calling,voice messaging,presence,group calling,call recording,transcription and summary.As fixed-line operators migrate their voice services to IP,and as mobile o

91、perators deploy voice-over-LTE technology,they have the possibility of adding new features and functionalities.Some have committed to the Google-backed Rich Communications Services(RCS)messaging standard,the prospects for which have been massively boosted by Apples decision to support it in its iOS

92、operating system.Microsoft,meanwhile,has been developing a suite of capabilities that enable CSPs to introduce“Teams-like”services to operators fixed and mobile customers for use over their existing devices.For example,it has been partnering with BT since 2021 to develop new voice calling services.B

93、ut what do these types of services have to do with AI?IP voice services do not use or rely on AI technology.But they do create a capability to enable the telecoms operator or the vendor supporting the operators new IP voice services to start offering new services that do use AI.This is achieved by u

94、sing GenAI foundational models to use the natural language in voice calls that customers make and receive to deliver new services.This natural language can relate to the calls that a customer makes and receives and which can be recorded,transcribed or translated.GenAI foundational models can use the

95、 natural language in voice calls that customers make and receive to deliver new services.For example,GenAI could be used to protect customers against unwanted calls a clear opportunity given the continued growth in scam and fraud calls.Microsoft Azure has built a product called Azure Operator Call P

96、rotection Preview which alerts customers via an SMS if the call they are making is likely to be a fraud or a scam.The service works by connecting the customers(IP)call to an Azure Communications Gateway platform,thereby integrating the customers landline or mobile voice network and the Call Protecti

97、on service running on the Azure platform.Concierge servicesConcierge is an emerging category of GenAI-enabled services that can be offered by any company that can collect and use data relating to a persons,or an organizations,personal or work life(or both).A concierge,in this context,is a virtual as

98、sistant that is designated to an individual person to provide them with recommendations,or which goes a stage further and actually automates different aspects of that persons day-to-day life.The more spoken or written content that a CSP can get access to,the better chance of it being able to deliver

99、 value in a concierge service.As such,if CSPs are to build such services they need to capture and act on the content of their customers voice and text communications.17inform.tmforum.orgIntegrating this content with other online tools for example,email and calendar services offered by providers such

100、 as Microsoft and Google allows the operator to offer services that go beyond mere advice and recommendations by automating scheduling and customer relations.Microsoft,for example,is building a suite of concierge services for small businesses in a range of different vertical sectors in partnership w

101、ith Norwood Systems and its CogVoice voicemail and voice services platform.Japanese telco Rakuten,meanwhile,is trialling a concierge service,Rakuten AI Assistant,for its mobile business and is leveraging data captured from its customers usage of other Rakuten services.It is unlikely that operators w

102、ill be able to charge a fee for concierge services in the consumer market.However,there are clearly opportunities in B2B,either as standalone services or as a service that is bundled into a high-end mobile subscription.That said,operators often struggle to demonstrate the value of dedicated mobile a

103、nd broadband price plans in the SME sector,and users are often content with using standard consumer packages.From IoT to AIoTThe term artificial intelligence of things(AIoT)has already been coined to describe new AI-enabled IoT services.While IoT essentially is about devices talking to each other vi

104、a the internet,AI helps these devices to learn from their data and their experiences.Vodafone IoT has already adopted the term Sight as a Sensor for a category of IoT applications that use still or video images to generate data that is then fed into an AI model to get results that help deliver insig

105、hts.For example,Vodafone has been running a trial with pest-control firm Rentokil that uses images and connected sensors to provide earlier signs of rodent infestations than would previously have been possible.A broader opportunity is opening up in the use of 5G connected CCTV cameras to help protec

106、t critical national infrastructure.A new class of video cameras is now able to function for as long as a year without the need to replace or recharge batteries.CSPs have long aspired to deliver greater value in IoT than merely connectivity,which generates less than 10%of total IoT revenues.By delive

107、ring the IoT device/sensor and analytics,in addition to connectivity,we estimate that CSPs have the chance to capture 50%or more of total solutions revenues.There are clearly opportunities in B2B concierge services,either standalone or bundled into a high-end mobile subscription.18inform.tmforum.org

108、CSPs will likely work with their security software and services partners or with hyperscalers to integrate AI into their cybersecurity services.(Cyber)security servicesManaged security services have emerged as an important and fast-growing line of business for many tier-one CSPs.Large B2B service pr

109、oviders such as Orange,which employs 3,000 cybersecurity experts,derive 2%or more of their total revenues from security services in B2B(Orange Business is targeting 1.3 billion in cybersecurity revenues in 2025).CSPs will most likely work with their existing security software and services partners o

110、r with hyperscalers to integrate AI into their cybersecurity services.AI can analyze vast amounts of data,detect patterns and predict future cyber threats.When it comes to responding to security threats,for example,AI can enable automated incident response and deliver managed services to address thr

111、eats without manual intervention.In the next section we look at how CSPs are working on large language models(LLM)and whether that presents commercial opportunities.19inform.tmforum.orgsection 3monetizing large language models20inform.tmforum.orgThe last two years has seen an explosion in new large

112、language models(LLMs)developed by both established tech players and start-ups.Operators are not significant players in the training of new LLMs.Most are using techniques such as fine-tuning to take existing models and inject them with their own data from customer operations,the network and other par

113、ts of the organization.But a handful of operators have committed to training their own models.They will use these LLMs for their own internal purposes,but part of their business model is also to generate revenues by making them available to third parties.We have identified three main approaches to t

114、he monetization of LLMs from these operators,and which are distinct from reselling GenAI services:Creating LLMs that are specific to the telecoms business and,most likely,in a specific domain such as customer operations or network Training an LLM that can then be tuned for specific market segments w

115、ithin that country Building an LLM dedicated to a specific language and where there is a belief/confidence that linguistic and cultural empathy and accuracy provides a unique selling point.Telco-specific LLMs The Global Telco AI Alliance comprising SK Telecom,Singtel,Softbank,e&and Deutsche Telekom

116、is developing multilingual LLMs using telecoms operator data.The early focus of the work from the alliance has been contact center operations,providing agents with real-time information during calls and post-call automation resolutions.But more recently,SK Telecom has been assessing the potential to

117、 use LLMs in the network and how to move beyond text-based output to multimodal content including visual and audio output based on unstructured data.Given that the initiatives of the alliance are designed to address issues specific to operator businesses,other CSPs could use its solutions as an alte

118、rnative to more generic LLMs.However,the alliance itself has not specifically stated that its aspirations lie beyond meeting the needs of its founding members.Another opportunity for the alliance would be to partner with business and operational support system(BSS/OSS)vendors.It could help both vend

119、or and operator communities address issues around complexity and interoperability when CSPs use vendor solutions based on different LLMs.BENCHMARKTM Forum|March 2024building anAI strategy:foundations in placetelcos put theAuthor:Mark Newman,Chief Analyst Editor:Ian Kemp,Managing EditorSponsored by:R

120、EPORTTM Forum|December 2023operators take their first stepsAuthor:Mark Newman,Chief Analyst Editor:Ian Kemp,Managing Editor Sponsored by:generativeAI:Read these reports to learn more about telcos AI and GenAI strategies:21inform.tmforum.orgTraining LLMs for specific market segments China Mobile is a

121、nother operator that has trained its own LLM,Jiutian.This has been developed both for its own internal usage and as a product and capability to generate new revenues.China Mobile is adapting Jiutian to build many different LLMs to address horizontal business functions such as customer care,telecoms-

122、specific LLMs including the network and IoT,and vertical market LLMs such as for healthcare,entertainment and smart cities.Unlike SK Telecom and the Global Telco AI Alliance,China Mobile has been quite explicit about the monetization of Jiutian being a key KPI by which its success will be judged.In

123、its first-half 2024 results presentation the company said it had deployed high-performance industry-specific LLMs in 40 sub-sectors:“In the first half of 2024,we have released 17 self-developed Jiutian industry-specific large models,implementing the applications of large-scale models across multiple

124、 industries such as public administration,IoT-enabled irrigation,and social general administration and governance.”The operator said it had launched 23 AI+products and 20 AI+DICT(digital,information and communications technology)industry applications.Japanese telco NTT has also developed a local lan

125、guage LLM called tzuzumi,specifically for Japanese businesses,which it will start offering in March 2025.In an effort to reduce the costs to businesses of training,inference and tuning,NTT has introduced lightweight versions of the model with parameters that are a fraction of the size of,for example

126、,OpenAIs LLMs.One of the verticals targeted by NTT is healthcare,and more specifically the analysis of medical data and drug development.In November 2024 NTT announced the general availability in Japan of tsuzumi on Microsofts Azure cloud computing platform through what it calls model-as-a-service(M

127、aaS).The pay-as-you-go pricing model is designed to make tsuzumi accessible to businesses looking to develop advanced AI capabilities at scale.“This model delivers efficient Japanese AI capabilities that excel in summarization,content generation and information retrieval that makes it easy for custo

128、mers to build innovative AI solutions,”said Stephen Boyle,Global Leader,GSI,ESI and Advisory Partners,at Microsoft at launch of the service.Building a language-specific LLMMost AI models to date are trained using English,and other languages are sometimes overlooked due to the lack of content availab

129、le for training purposes.For operators in countries with languages for which there is a lack of available content this presents both a challenge and an opportunity.Korea Telecom is building a local-language LLM for Thai technology group Jasmine Technology,while French telecoms group Orange is partne

130、ring with OpenAI and Meta to train LLMs with natural language content in several west African countries.Natural language is the most compelling GenAI-enabled customer service interface in countries with low levels of literacy.Training or fine-tuning LLMs in these languages can help reduce the risk t

131、hat societies miss out on the benefits of new AI services and applications.Training LLMs with natural language content can help reduce the risk that societies miss out on the benefits of new AI services and applications.22inform.tmforum.orgThe Global Telco AI Alliance,meanwhile,is training LLMs in a

132、 range of languages including Korean,English,German,Arabic and Bahasa.At DTW24-Ignite in June 2024 the alliance members shared a stage with their hyperscaler partners to explain their LLM tailoring strategies.“Were probably working with about 20 different sovereign governments right now to build spe

133、cific LLMs for that country or region,”said Sameer Vuyyuru,Director,Global Business Development for Telecom,Amazon Web Services,in an interview with TM Forums Inform.In December 2024 Deutsche Telekom launched an open-source AI LLM that supports all 24 official European Union languages.The operator s

134、ays it is the first commercial offering using OpenGPT-Xs Teuken 7B model and is designed to provide secure digital sovereignty solutions for businesses and public authorities.“Providing Teuken as an open-source model has multiple advantages,”said Dr.Nicolas Flores-Herr,Project Lead at Fraunhofer IAI

135、S,one of the partners in the consortium that developed the model.“Companies can tailor the model to their needs,developing specialized applications.They can also choose to run the model locally on their infrastructure or with a trusted cloud provider of their choice.This allows sensitive data to rem

136、ain within the organization.”Getting a return on LLM investmentsBut it is difficult to see how CSPs can generate a return on investment in LLMs based on external monetization using standard token-based charging principles.“Theres some evidence that it is a race to the bottom because the price per to

137、ken of generated LLM output has dropped 100 x in the last year,which is much faster than Moores law,”said Marc Andreessen,general partner of venture capital firm Andreessen Horowitz,at a tech event in October.As such,CSPs will need to find other commercial models or justify investment based on impro

138、vements in productivity or customer experience.Rather than building their own models some CSPs are seeking to collaborate with hyperscalers and other leading developers of LLMs.Korea Telecom,for example,is partnering with Microsoft and plans to develop customized AI models including OpenAIs GPT-4.o

139、and Microsofts Phi 3.5 small language models.These AI models will be co-developed and rolled out to KTs customer services such as chatbots and industry-specific AI solutions for B2B sectors.As part of the deal,Microsoft will pay Korea Telecom to use its cloud infrastructure.The five-year multi-billi

140、on dollar deal aims to drive AI innovation.It includes the establishment of an AI-focused company to deliver AI transformation to enterprises in Korea,joint R&D through a co-innovation center,development of industry-specific AI models,development of sovereign cloud solutions,and upskilling of KTs wo

141、rkforce.In November KT also announced it would restructure to pursue its focus on AICT services.In the next section we look at how CSPs are starting to use platforms to deliver AI products and services.Author:Dean Ramsay,Principal Analyst Editor:Ian Kemp,Managing EditorSponsored by:labsknowledgecode

142、+frameworkstraining+accreditationimplementing GenAIhow to set up and maintain large language modelsRead our e-book to find out more about how to set up and maintain LLMs:using platforms to deliver products and servicessection 423inform.tmforum.orgThe term platform is used a lot in CSPs statements ab

143、out their strategies for monetizing AI.Allied to this are many different products,services and value propositions.For the purpose of this piece of research we define platform simply as a means for CSPs to deliver different AI-infused products and services.Platforms represent a means by which telcos

144、can insert themselves between the fast-moving,chaotic AI product and services landscape and customers who may not have the capability or desire to buy directly from new,untried providers.In many cases CSPs platforms have been developed first for internal use cases and applications and only subsequen

145、tly exploited for external users.This is the case with Live Intelligence services,launched by Orange Business at the end of November 2024 and which was road tested by 50,000 of its own employees for a year.A multi-LLM,GenAI software-as-a-service offering,it aims to help employees of enterprises and

146、local authorities improve operational efficiency through the safe and managed use of GenAI.It does this by providing access to a library of pre-set prompts to address common business GenAI use cases,such as analyzing a document,extracting important information from an email chain,writing meeting min

147、utes or drafting an agenda.Orange Business sees data security and data sovereignty as key selling points of Live Intelligence,which initially launched in France but will be extended to other areas of Europe.Not only does it provide an alternative to uncontrolled adoption by employees of free online

148、24inform.tmforum.orgGenAI tools that could provoke data breaches,but the data it uses will also remain in Europe.And it comes with a dashboard to monitor how employees are using GenAI,including the type of LLM.By inserting a platform between its own users and the LLM itself,an organization can ensur

149、e that it retains the knowledge captured from inputs and outputs from different models.This can help efforts to fine-tune different models and,potentially,make more intelligent use of different LLMs to manage costs.There is also an opportunity for CSPs to play the role of broker,offering bundles of

150、services from different vendors at price points that match their usage.Spanning the AI value chainMany CSPs are trying to use AI to enhance their 5G value propositions and help build compelling B2B services.JioBrain is a GenAI-enabled platform integrated with 5G to enable advanced network management

151、 and service delivery capabilities.It has been developed by Jio Platforms,a subsidiary of mobile operator Reliance Jio,and is initially being deployed for internal use cases and for other Reliance companies.The plan is then to offer it to other companies.There is an opportunity for CSPs to play the

152、role of broker,offering bundles of services from different vendors at price points that match their usage.25inform.tmforum.orgSingtel is another company with big AI ambitions.In October it announced RE:AI,a group of products and services that aims to make AI available to enterprises.Singtel describe

153、s it as an AI development and deployment platform.One of RE:AIs offers is cloud-based enterprise AI capabilities delivered over 5G,using Paragon,its in-house 5G service orchestration platform for enterprises.According to Singtel it combines compute infrastructure such as graphics processing units(GP

154、Us)and storage,workspaces and tools with 5G,fixed and quantum-safe networks.“Many enterprises and public sector customers have shown keen interest to bring AI into their operations,”said Bill Chang,CEO of Singtel Digital InfraCo,in a launch statement.“However,the high costs and long lead times for G

155、PUs,the need for special environments to host them due to their intense energy utilization,the complexity of AI technology and the lack of talent are key friction points in their respective journeys.”Japanese telecoms group KDDI has launched a“business platform”called WakonX(pronounced Wakon-Cross),

156、which brings together its networks,computing and data resources and its AI and vertical market knowledge into a single proposition for a monthly subscription.WakonX has a strong focus on IoT and 5G.It is developing“standardized tools”and solutions across multiple vertical sectors including mobility/

157、automotive,retail,logistics and entertainment.KDDI says AI is central to the three layers of its WakonX platform:network,data and vertical.There is a platform element to the AI factory strategies that are being pursued by a number of CSPs.Collaborating to develop new AI use casesThere is also a plat

158、form element to the AI factory strategies that are being pursued by a number of CSPs in collaboration with Nvidia and which we explored in section 1.At the launch of its AI factory and the onboarding of its first customer in November,Scandinavian telecoms group Telenor said it would serve as“a colla

159、borative platform”combining the operators infrastructure and expertise with partner companies to accelerate the adoption of AI across industries.As we have seen already in this report in developments from Korea Telecom and the Global Telco AI Alliance,for example collaboration is key to many platfor

160、m developments.Two years ago Vodafone announced the launch of its AI Booster AI/machine learning platform.It uses Google technology to enable and automate the deployment of AI use cases such as optimizing customer experiences and product recommendations.Built entirely on Google Cloud,using the Verte

161、x AI development platform for building GenAI applications and machine learning models,AI Booster is a fully managed cloud-native platform that integrates with Vodafones Neuron platform,a data ocean also built on Google Cloud.One of the options now open to Vodafone is to use AI Booster to build new p

162、roducts using its own network,customer and operations data.In the next section we look at how CSPs are placed to build managed AI services businesses.building an AI managed services strategysection 526inform.tmforum.orgManaged services are an important line of business for any operator that has a la

163、rge established B2B function.Most managed services revenue comes from network services,with large national and multinational corporations using a mixture of private and public infrastructure.But the managed networked services business is stagnating as enterprises transition to cloud and IP connectiv

164、ity.As such,operators are relying on revenues from new managed services such as cloud and security to deliver topline growth.These businesses typically are growing at 10%-15%per year on a global basis.Managed AI services represent another potentially high-growth opportunity.Gartner predicts that 80%

165、of enterprises will have used GenAI APIs or deployed GenAI applications by 2026,but they face several challenges in deploying this technology.High entry costs,technical complexity and concerns about data leaks are significant barriers to adoption,for example.Delivering professional services to these

166、 companies is already a fiercely competitive market dominated by systems integrators,specialist managed service providers and technology vendors.Telcos will need to figure out which solutions and which market segments offer them the best chances of success.27inform.tmforum.orgCSPs need to address a

167、range of issues as they consider building a managed services capability in data and AI:Are they building the necessary internal skills and tools to make them credible providers of AI services?Do they have the consulting skills to deliver such services?Which specific services from a solution perspect

168、ive should they seek to develop?Which market segments offer the best opportunities?Who should they partner with?Do they need to make acquisitions to establish a presence and a capability?How do they define success in terms of revenues,EBITDA margin and profits?Large B2B operators have already built

169、teams of experts in different fields to become credible providers of managed and integration services.For example,Orange has 3,000 cybersecurity experts worldwide.But the race for AI talent has heated up in the past two years as a result of generative AI,and operators are using a combination of retr

170、aining,recruitment and partnering with systems integrators or hyperscalers to bring in the right skills.High entry costs,technical complexity and concerns about data leaks are significant barriers to GenAI adoption for enterprises.28inform.tmforum.orgIn the previous section,for example,we saw that p

171、art of KTs major new partnership deal with Microsoft involved upskilling the operators employees in AI technologies and services.A TM Forum survey of 40 CSPs conducted in the first half of this year revealed that cybersecurity and AI/machine learning are the skills in most demand across the telecoms

172、 sector(see graphic).“CSPs can live with outsourcing at the engineer level,but at the architect level its insourcing and especially for these key areas of cloud,AI,cybersecurity,”explained Ashish Yadav,Senior Director,Engineering,at Capgemini,in our report Finding skills for the future:inside the te

173、lco talent revolution.The main requirement for skills comes from the internal technology functions rather than the line of business.Becoming AI-enabled is now a strategic imperative for most CSPs,and in the skills survey it was cited as the most important reason for insourcing talent.Conversely,inso

174、urcing talent to help drive revenue growth be it for data monetization,reselling platform capabilities or for managed services ranked as the least important business drivers(see graphic on the next page).An inside out modelIf CSPs are to build a managed(AI)services business they will first need to b

175、uild skills,products and systems for their own internal requirements.For example,some of the CSPs that are most advanced in their adoption of GenAI have built platforms for their own internal teams that allow them to use many different LLMs and a range of different techniques.These include retrieval

176、 augmented generation(RAG)a way to improve the output of an LLM and Which skills are in highest demand?TM Forum,202429inform.tmforum.orgprompt engineering the process of refining LLMs with specific prompts and recommended outputs to optimize results.Could CSPs then make this platform capability avai

177、lable to their own customers?This is precisely how Orange has gone about commercializing its Live Intelligence multi-LLM GenAI platform and service that was launched at the end of November 2024 as we saw in section 4.“The solution was originally conceived by Orange innovation teams to transform its

178、own work practices and increase its efficiency by providing access to the latest AI solutions,”the company said at launch.Orange Business has ambitious targets to build AI capabilities.Its intention is to have 3,000 AI experts within the next five years.Swisscom is another telecoms group which belie

179、ves that building its own internal skills and capabilities is a requisite for setting up an AI managed services practice.There was some surprise from industry commentators when the company entered into a strategic partnership with Nvidia at the start of this year and committed to investing SFr100 mi

180、llion(about$113 million)in an AI Factory to deliver services to customers in Switzerland and Italy.Isa Mueller-Wegner,Head of Strategy and Business Development at Swisscom,stresses the importance of seeing the partnership(and the investment)in the context of Swisscoms role as the largest provider of

181、 IT services in Switzerland.“The Nvidia partnership is really more about how we help our customers make the most of AI,”she says.“And the context here is that lots of people tell us,hey,look,this all sounds very exciting,but Im nervous about where my data is going,Im nervous about knowing what data

182、my models are trained on,and Im nervous about where my information is held.And wed really quite like you to come and help navigate this,because Im not comfortable sticking my data into a machine which potentially goes way outside of our sovereign areas.”In the first commercial developments from that

183、 partnership,Swisscom in November 2024 launched its Swiss AI Platform,which provides AI-focused sovereign data storage and processing in data centers in Switzerland.Targeting enterprise users in verticals including finance and the public sector,customers can source AI services CSPs reasons for insou

184、rcing talentTM Forum,202430inform.tmforum.orgfrom Swisscom ranging from consulting through to the development of AI applications,implementation and operation.The platform is based on what it calls the Nvidia SuperPOD System(full-stack supercomputers).Initial applications include GPU-as-a-service;an

185、AI work hub designed for data scientists and machine learning engineers;and,from spring 2025,a studio to give companies access to GenAI services via API interfaces on the Swiss AI Platform,which they can use to create their own AI solutions.Swisscom says it has 400 AI and data specialists working al

186、ongside Nvidia experts to provide tailored AI solutions.Hyperscaler partnerships Operators have been talking for many years about their desire to build partnerships with hyperscale service providers that go beyond simple buyer-seller relationships.Managed AI services offer a good opportunity to buil

187、d such relationships.The best opportunities for both CSPs and hyperscalers are likely to be in the SME market where operators have already built trusted customer relationships,where large professional services firms do not generally have a presence,and where there is often a lack of knowledge and sk

188、ills about how to leverage new technologies such as AI.But to address this market effectively and profitably,CSPs will need to scale products down to make them affordable and easy to use.Orange certainly has its eyes firmly on the SME market.“Live Intelligence aims to accelerate the managed adoption

189、 of GenAI for all businesses,including SMEs that represent the core of the European economy,”said the operator at launch of the software-as-a-service offering.In the final section we summarize the report findings and give some recommendations for CSPs looking to leverage AI for service and value cre

190、ation.REPORTAuthor:Dawn Bushaus,Contributing AnalystEditor:Ian Kemp,Managing EditorSponsored by:inside thetelco talentrevolutionfinding skills for the future:TM Forum|May 2024Read our report to find out more about the skills landscape in the telecoms sector,including the potential impact of AI and a

191、utomation on skills and jobs:key findings and recommendationssection 631inform.tmforum.orgOperators are at the very start of their journeys to build products,services and solutions that are enabled by,and fused with,artificial intelligence.As such it is too early to identify strategies that have bee

192、n successful and those that have failed.However,in this section we have 10 recommendations for CSPs as they consider how to build and execute a strategy for generating new product and service revenues from AI.These recommendations are based on learnings from CSPs attempts to enter the cloud business

193、 10-15 years ago and from a broader understanding of their capabilities in the ICT services market.Build your internal capabilities firstIf CSPs want to persuade enterprises they are capable of helping them to use AI in their businesses they will first need to demonstrate that they have built,and ar

194、e using,their own internal platforms and capabilities.For example,many businesses want to use generative AI to transform their customer service operations to deliver a better,customized experience,ideally at a lower cost.CSPs will only be considered as partners for such projects if they can demonstr

195、ate that they have successfully used GenAI to improve customer service for their own customers.32inform.tmforum.orgCreate roles and accountabilities for AI value creationIn many CSPs responsibility for AI sits either within the technology function or is divided between many different functions with

196、nobody seeing it as one of their priorities.Furthermore,the primary benefit of AI,to date,is seen as a productivity and efficiency tool rather than a driver of business value.CSPs either need a CXO role which is dedicated to realizing the benefits of AI for both efficiency or value creation and/or t

197、hey need a function within the commercial or product organization which is fully focused on AI monetization.CSPs will only be considered as partners if they can demonstrate that they have successfully used GenAI to improve customer service for their own customers.33inform.tmforum.orgTo be successful

198、 in the AI era,and in value creation,effective partnering will be a necessity not a choice.Improve your partnership capabilitiesOperators have been talking up the importance of partnering for many years,but the evidence to date is that they are not truly serious or simply do not have the right cultu

199、re or skills to partner effectively.But to be successful in the AI era,and in value creation,effective partnering will be a necessity not a choice.Partnering will be needed to bring new third-party products,services and applications to market and to bolster CSP skills in delivering managed AI servic

200、es to B2B customers.That collaboration will be possible across the entire AI value chain:from infrastructure to platforms,to AI models to AIOps services and tools,and to products for consumers and business customers.Pick your battlesIn recent years and more specifically since the launch of GenAI ser

201、vices at the end of 2022 large professional services companies have built formidable teams of data and AI experts.Competing with these companies is likely to be a thankless task for CSPs.However,these companies are only targeting the very high end of the large business,multinational and public secto

202、rs.This leaves plenty of opportunity for CSPs to compete in the SME space a market they already serve but which might not consider operators to be providers of AI services.However,operators reputations as reliable service providers could count in their favor if they can present themselves as trustwo

203、rthy partners to navigate the complexities and risks of the AI technology sector.Become a retailer of IT products and servicesThe coming years are going to see an avalanche of new AI products and services.Many of these will come from leading digital brands and service providers.Operators will have t

204、he opportunity to resell these products,but are unlikely to secure a large profit margin.However,there may be opportunities to resell products from lesser known companies and brands that are prepared to offer them bigger profit margins because of their ability to reach deep into consumer and B2B mar

205、ket segments.Leverage your key assets:dataIt is a given that CSPs will leverage their own data from the network,from their customers and from functions across the organization to build AI use cases.But these use cases will mainly be for internal purposes rather than for customer products and service

206、s.There may be opportunities to use this data on an aggregated basis and without breaching rules relating to security and privacy to build either new products or to enhance existing ones.34inform.tmforum.orgAs national service providers,CSPs can give businesses assurance that their data will be kept

207、 within national borders.Leverage your key assets:networksIt is still too early to know precisely where AI workloads will fall most heavily on telecoms networks.It will be a function of where AI models run and their requirements for low latency.But if we assume that AI has the kind of impact on busi

208、ness and society that we expect,it is a reasonable assumption that it will change telecoms network traffic flows.If this is the case,opportunities will inevitably arise for CSPs to help businesses AI users or AI service providers to achieve better results.Leverage your key assets:trust,reliability,n

209、ationalityCSPs traditional reluctance to innovate,to take risks and to extend their product reach outside of their home markets could work in their favor in the emerging AI economy.Enterprises will naturally be cautious when it comes to deciding where to store their data,what models their data is tr

210、ained on and whether they can rely on the results.As national service providers,CSPs can give businesses assurance that the data will be kept within national borders.When it comes to providing assurances about the accuracy from foundation models,CSPs can offer customer access to the same platforms a

211、nd tools that they use internally.Rethink your edge strategyFive years ago edge computing in combination with 5G connectivity was seen as an attractive new business opportunity for mobile operators.But the market has failed to gain traction.Many operators have launched so-called public MEC(multi-acc

212、ess edge computing)services,but demand for a shared computing platform located close to business premises has never materialized.Some CSPs are now dusting off their edge computing strategies in the hope that AI will create a demand for edge processing.But is the result going to be any different seco

213、nd time around,or do operators need to rethink whether there is genuine business value in 5G+edge+AI?Is edge computing satisfying a genuine business need,or is it just a technology looking for a market?Dont get swept up in the AI infrastructure arms raceThere is a huge AI arms race taking place acro

214、ss the world as tech companies and data center firms seek to build the next generation of cloud computing infrastructure that is optimized for AI.But should CSPs be participating?If a CSP has been successful in operating a data center or a public cloud business in the past,it may have good reason to

215、 believe that it can do so in the AI era.But for those CSPs investing in cloud and data center infrastructure for the first time,the risk is that they use technology which will become outmoded before they are able to create a return on their investment.35inform.tmforum.orgUnlocking AI-powered telco

216、product and service creationAdditional feature by CerillionWhen it comes to creating new products or promotions,most telcos still face great difficulty in translating marketing ideas into the corresponding BSS/OSS configuration.A new idea might struggle to make it off the drawing board simply becaus

217、e everyone assumes that it will take too long to build,test and launch,stifled by the perception of complex implementation,lengthy development cycles,and the inertia of legacy systems.As digital transformation continues,CSPs are looking to harness AI for ways to open up new business use cases and re

218、venue streams,optimise operations and deliver personalised services.Facing increasing pressure from both traditional competitors and more agile digital-native providers,AI has the potential to be an accelerator for CSPs at risk of falling behind in competitiveness and innovation.No longer are telcos

219、 constrained by outdated processes and rigid systems AI is helping organisations to automate business processes,overcome technological limitations,and reimagine their whole operational ecosystem.There is now a huge opportunity for telcos to integrate AI-infused products and services into their offer

220、ing portfolios.Enhanced network servicesCSPs are increasingly deploying AI-powered solutions for 5G network optimisation and management,analysing network performance data to predict potential failures in real-time,and optimise network services.New network capabilities,such as network slicing and low

221、-latency services,can be further enhanced with AI to create more differentiated services and custom solutions for specific industry use cases.Whether youre a remote worker requiring consistent video conference quality,a gamer demanding ultra-low latency,or a smart city infrastructure manager needing

222、 reliable IoT connectivity,your network experience can be precisely tailored to your requirements using AI.By shifting from an internal efficiency perspective to a focus on optimisation of the customer experience,AI can redefine how people interact with and rely on their communications services.Cust

223、omer experience and service deliveryModern GenAI solutions are reinventing the digital experience,speeding up how CSPs build and launch new products and services,and how customers find,buy,personalise and consume them.Powered by AI,CSPs can:Develop hyper-personalised service packages,dynamic pricing

224、 and targeted marketing strategies Rapidly create new promotional offers,launch new bundles and add-ons,including third-party products and services via marketplaces Predict and proactively address potential issues with the products and services they are sellingNew network capabilities,such as networ

225、k slicing and low-latency services,can be further enhanced with AI to create more differentiated services.36inform.tmforum.orgUnlocking AI-powered telco product and service creationAdditional feature by CerillionCSPs can take the lead in offering AI-as-a-Service solutions to their B2B customers,comb

226、ining AI technology partnerships with their own cloud capabilities.Create more intuitive and responsive customer engagement platforms to track customer sentiment and enhance overall satisfaction Automate the complete Lead-to-Cash process with GenAI providing the decision-making on what are typically

227、 manual steps,as well as identifying anomalous and unusual actions before theyre executedUnderpinning this journey is the discovery of new products and services through natural language interactions with AI-powered virtual assistants and chatbots.These might be initiated by the customer during the s

228、ales process “please find the best broadband package to meet my needs”;or suggested by the virtual assistant during a customer service request “I noticed that your contract expires next month,we have some special offers available if you renew today”CSPs can now leverage AI to analyse customer behavi

229、our patterns,including their engagement across all touchpoints and consumption data,to create hyper-personalised service offerings and custom tariffs.With unprecedented precision,based on personal usage patterns and lifestyle contexts,these models understand needs and preferences,suggesting products

230、 accordingly using the customers requirements and situational context,and opening the door to more cross-selling and upselling opportunities,resulting in improved satisfaction scores.AI-as-a-Service With AI embedded into the products and services they sell,as well as their own operations,the next va

231、nguard for CSPs is AI-as-a-Service,leveraging their trusted brands to expand beyond traditional connectivity services.CSPs can take the lead in offering AI-as-a-Service solutions to their B2B customers,combining AI technology partnerships with their own cloud capabilities,data security and BSS/OSS a

232、ssets to help businesses to quickly launch and monetise their own AI-infused products and services.For example:Computer vision solutions Natural language processing Predictive maintenance Data analytics platforms Fraud prevention and cybersecurityWe can expect to see even more innovative application

233、s and business models emerge,further cementing AIs role as a critical driver of growth and innovation in the telecommunications sector.The CSPs that will thrive in this AI-driven future are those that can effectively balance innovation with practical implementation,while maintaining a strong focus o

234、n customer value and operational excellence.37inform.tmforum.orgUnlocking AI-powered telco product and service creationAdditional feature by CerillionHow Cerillion can helpToday,AI has provided a target rich environment and many possible use cases,but all requiring work to deliver the predicted bene

235、fits.Few CSPs have the luxury of dedicated resources assigned to assess and deploy AI,either at a companywide level or within the departments and teams.As a technology partner,Cerillion has been at the forefront of AI adoption in the telecoms industry since late 2022,both as an early adopter in the

236、BSS/OSS vendor community,embedding GenAI capabilities throughout our product suite,and by deploying these tools and approaches within the wider company.Our GenAI-enabled Enterprise Product Catalogue allows instant translation of ideas into fully configured services,automatically creating new product

237、s,packages and workflows based on natural language and image recognition,dramatically simplifying the product development lifecycle and democratising access to the product catalogue.Natural language text and voice processing means users simply describe what they want to create in their own words.Wit

238、h image recognition,users can also upload pictures and diagrams from which new products or packages will be automatically created,cutting the product development cycle by as much as 95%.38inform.tmforum.orgUnlocking AI-powered telco product and service creationAdditional feature by CerillionAnd by c

239、ombining image recognition and natural language instructions,users can further modify or accelerate their product ideas e.g.“copy this triple-play package but make the monthly fee 10%cheaper”.The large language model(LLM)interprets these instructions and creates the corresponding product configurati

240、on in a matter of seconds,ready for validation,testing and launch,completely transforming how CSPs can design and build new products and services.These capabilities can also be extended to key business processes including order management,service orchestration and network service delivery.Users can

241、create new workflows straight from pictures or swim-lane diagrams,reducing provisioning errors on complex products,and bringing down time-to-market considerably.All this is accomplished through a generic architecture,that allows for integration with multiple LLMs via API,including GPT,Claude and Gem

242、ini,based on business needs.Whats next for telcos?Artificial intelligence is increasingly determining competitive advantage in the industry,and how communications services can be delivered and experienced.As LLMs continue to evolve,we can anticipate even more sophisticated,adaptable and powerful AI-

243、infused products and services,accompanied by a lower cost of entry for using these new tools.CSPs have an unprecedented opportunity to create more intelligent,responsive and value-driven ecosystems,as the technologies underpinning these systems mature.Successful CSPs will be those that dont view AI

244、as a standalone technology,but as a core component of their digital transformation strategy one that will speed up how new products and services are built and launched,creating new opportunities for value while addressing longstanding operational challenges.39inform.tmforum.orgUnlocking AI-powered t

245、elco product and service creationAdditional feature by CerillionAbout CerillionCerillion is a leading provider of business and operations support systems(BSS/OSS)with 25+years experience delivering solutions to mobile,fixed,cable and multi-service communications providers worldwide.With 80+customers

246、 in 45+countries,Cerillion has a proven track record of delivering cost effective solutions to the billing,charging and CRM challenges of today and tomorrow.We combine leading edge products with highly skilled and experienced staff,to provide long-term solutions to your business challenges.From full

247、y integrated systems to managed services and SaaS,we offer a range of approaches and business models to suit your needs now and in the Furthermore,with the speed of evolution of the open source LLMs,we expect that these will soon be sufficiently powerful for CSPs to easily run their own private LLMs

248、 to incorporate customer and internal data,while ensuring that data sovereignty is maintained.By removing the dependency on IT and technically minded teams,GenAI can streamline the product creation process and put business users in full control of the product lifecycle,empowering them to easily buil

249、d and configure new offerings,accelerating time-to-market and enhancing the user experience.Yet without significant investment in modernising core systems and acquiring or training the necessary talent,CSPs will face challenges integrating AI into their legacy systems;CSPs can address this through p

250、artnering with AI-focused companies to accelerate adoption and bridge the technology and talent gaps that currently impede their digital transformation strategies.Cerillion brings productised capability and real-world industry experience of how these tools can turn ideas into revenue and cost saving

251、s.Incorporating GenAI into the process of building and launching new products and services is a practical first step towards wider AI integration within CSP businesses,offering measurable outcomes and tangible benefits,rather than buzzword compliance.Contact us now at: 40inform.tmforum.orgThe New Im

252、perative:Transforming Telecommunications for the AI EraAdditional feature by PegasystemsWith the telecommunications industry investing over$140 billion annually in R&D,the transformation to AI-enabled services represents a critical pivot point for future growth.Despite powering the digital revolutio

253、n that has transformed virtually every sector of the global economy,Communications Service Providers(CSPs)face an existential challenge:their own business models have failed to keep pace with the very innovation they enable.While Over-the-Top(OTT)services continue to thrive on the infrastructure CSP

254、s have built and maintained,the industrys traditional revenue streams show a persistent decline.The Productivity ParadoxThe root of this paradox lies not in a lack of technological capability,but in how that technology is deployed and orchestrated.With 60-80%of IT budgets locked in legacy systems an

255、d siloed AI implementations,CSPs find themselves in a situation that echoes economist Robert Solows famous observation about the computer age:revolutionary technology is everywhere except in the productivity statistics.Traditional approaches of ripping and replacing systems have consistently fallen

256、short,with only 56%of transformation programs achieving their intended goals.The complexity of migrating to modern platforms has been consistently underestimated,leading to 18-24-month roadmaps that deliver only basic capabilities.Meanwhile,approximately 70%of CSPs investment capacity remains tied u

257、p in maintaining network infrastructure,servicing debt,and paying dividends.The industrys transformation efforts have often resulted in fragmented solutions that create shadow silos,leaving organizations struggling with disconnected systems and disjointed customer experiences.41inform.tmforum.orgThe

258、 New Imperative:Transforming Telecommunications for the AI EraAdditional feature by PegasystemsWith Orchestrated AI,the industry is set to return to growth across new AI-infused products and services and B2B managed services offerings.With Orchestrated AI,the industry is set to return to growth acro

259、ss new AI-infused products and services and B2B managed services offerings.To achieve this growth,CSPs must change how they think about introducing new technologies into their environments and focus on creating bridges,not silos.The Path Forward:Beyond Traditional TransformationThe path forward requ

260、ires a new paradigm that enables CSPs to break free from decades-old cycles of incremental improvement and embrace a model of continuous innovation.This new model is composed of three key elements:1.Orchestrated AI Integration Converting multiple AI silos across predictive,adaptive,generative,and ag

261、entic capabilities into a unified system Focusing on meaningful business outcomes rather than isolated technological achievements Enabling real-time decision-making and automated optimization2.Partner Collaboration for Growth Creating seamless integration pathways with internal and external partners

262、 Building standardized APIs and microservices architectures Establishing secure and scalable data-sharing frameworks3.Rapid Value Realization Delivering measurable ROI within 3-6 months Implementing modular solutions that can be rapidly deployed and scaled Creating feedback loops for continuous impr

263、ovementThrough this comprehensive approach to transformation,CSPs can finally break free from traditional constraints and create new paradigms of value creation in the digital age.42inform.tmforum.orgThe New Imperative:Transforming Telecommunications for the AI EraAdditional feature by PegasystemsWe

264、ve achieved the highest scores in both Current Offering and Strategy categories for Real-Time Interaction Management(RTIM),with our Customer Decision Hub.Pegas Proven RecordPegas leadership in the communications industry,the worlds largest 10 Telcos,is continually strengthened by the capabilities de

265、livered by the platform.As recognized in Forresters latest Wave reports,Pega continues to set the standard for enterprise transformation.Weve achieved the highest scores in both Current Offering and Strategy categories for Real-Time Interaction Management(RTIM),with our Customer Decision Hub praised

266、 as the centralized brain orchestrating billions of real-time interactions across channels.In addition,Forrester just released their Wave report for Task-Centric Automation Software(Q4 2024),and Pega has been recognized as a powerhouse in the task-centric process automation market.Weve achieved the

267、highest scores of all vendors evaluated in both the Strength of Offering and Strength of Strategy categories,validating our leadership in delivering sophisticated automation solutions specifically designed for communications service providers.Orchestrated AI is achieved when real-time customer inter

268、actions can be resolved in the back office by the best AI-powered automation software in the industry.Our industry leadership is further demonstrated in our groundbreaking Ignite 25 Catalyst project,Precognitive Network Orchestration,which promises to revolutionize network operations through:A reusa

269、ble CSP-wide framework delivering beyond 30%EBITDA improvement Quantum leaps in network orchestration through AI-powered transformation Integration of predictive,adaptive,generative,and agentic AI with Pegas Blueprint to lay the fou2ndation for Level 4&5 autonomous network maturity An orchestrated A

270、I approach combining Pegas capabilities with third-party AI for precognitive autonomyFurthermore,our commitment to innovation is evident in our recent TM Forum Catalyst award-winning projects:1.GenAI Genie Redefines CX(M24.0.634)-Focused on increasing customer lifetime value through generative AI,br

271、idging gaps between the customer,customer service,and network operations43inform.tmforum.orgThe New Imperative:Transforming Telecommunications for the AI EraAdditional feature by Pegasystems2.GenAI Hyper-personalized Customer Experience(C23.0.497)-Delivering award-winning results:Up to 35%increase i

272、n customer lifetime value Revenue growth up to 24%20%reduction in churn through AI-powered personalizationThe telecommunications industry cannot afford to wait for the future it must create it.By adopting an agile,intelligence-driven approach to transformation,CSPs can:Accelerate product launches fr

273、om months to weeks Turn partnerships into immediate growth opportunities Build sustainable cycles of innovation Ensure long-term competitiveness in an increasingly dynamic marketThe transition to an autonomous enterprise powered by orchestrated AI represents a strategic evolution rather than a techn

274、ological leap.This transformation enables:Intelligent Decision Distribution:AI handles routine,data-driven decisions at scale while human expertise focuses on strategic,non-routine decisions Strategic Alignment Through AI:A unified AI engine coordinates operations,service,and engagement,optimizing e

275、very transaction for strategic outcomes AI-Powered Operational Excellence:Automated workflows execute in seconds,driving hyper-personalization and increased productivity within existing resourcesConclusion:Pega is your partner for growth.With Pega,you will deploy the foundation and prove Pegas capab

276、ilities within 90 days;configure high-value use cases in 90 days;and scale over the next 18 months.This iterative approach enables you to address challenges with data privacy,legacy system integration/replacement,and workforce adaptation while achieving incremental value along the way.The telecommun

277、ications industry stands at a crossroads.The path forward requires more than technological advancement it demands a fundamental transformation in how CSPs operate,innovate,and deliver value.By embracing the principles of the autonomous enterprise and implementing orchestrated AI solutions,CSPs can b

278、reak free from traditional constraints and create new paradigms of value creation in the digital age.About PegasystemsPega is The Enterprise Transformation Company that helps organizations Build for Change with enterprise AI decisioning and workflow automation.Many of the worlds most influential bus

279、inesses rely on our platform to solve their most pressing challenges,from personalizing engagement to automating service to streamlining operations.Since 1983,weve built our scalable and flexible architecture to help enterprises meet todays customer demands while continuously transforming for tomorr

280、ow.For more information,please visit us at 2024 Pegasystems,Inc.All rights reserved.All trademarks are the property of their respective owners.44inform.tmforum.orgPaving the way to new revenue with an open and scalable AI platformAdditional feature by Red HatAuthor:Volker Tegtmeyer,Principal Product

281、 Marketing Manager,Telecommunications,Red HatAs communications service providers(CSPs)look to harness the power of AI to develop new products and services for their customers,they face multiple challenges.Addressing them starts with a comprehensive AI strategy that encompasses applications,platform

282、and organizational change.Most large telcos are already using hundreds of AI models across their businesses,with new generative AI(GenAI)among the most promising.To avoid creating new and disconnected technology stacks,a holistic view of these applications and use cases is necessary to understand ho

283、w various departments are using them.Starting on the right foot with a vision and strategy for AI is critical to addressing the three key challenges operators are facing:Ensuring independence from closed AI platform providers so that service providers can tap into a broad ecosystem of partners and c

284、apitalize on their own unique ability to deliver network-as-a-service(NaaS)Developing a robust data foundation and strategy that can improve intelligent automation and operational efficiency Integrating new AI technologies and applications with a vast number of legacy processes and systems.1.Maintai

285、ning control over a flexible AI platformThe most critical challenge to address is establishing independence from proprietary AI platform providers.CSPs need to be able to train AI models on their own data(see sidebar on the next page)and leverage the integration of AI services from cloud providers a

286、s well as open-source models.And they need to be able to train,deploy,run and scale these models anywhere from on-premises locations to edge environments to all types of clouds including public,private,hybrid and sovereign.Such an approach balances data security and sovereignty,provides the right le

287、vel of compute power needed for simple and advanced AI models,and therefore helps manage the costs of using AI.It also empowers telcos to build a more competitive ecosystem of solutions tailored to their specific business and technical needs.For example,a startup AI vendor might offer new methods of

288、 hardware acceleration,which the operator could then use to create innovative services for enterprise customers.2.Leveraging data and increasing automationAnother challenge operators face in becoming AI-driven lies in the development of a comprehensive data strategy.AI thrives on data,and CSPs posse

289、ss vast amounts of unique data,both telco and non-telco related.However,the data is often stored in individual databases and spread out across different parts of the organization,making it difficult to access,leverage and monetize.To avoid creating new and disconnected technology stacks,a holistic v

290、iew of AI applications and use cases is necessary.45inform.tmforum.orgPaving the way to new revenue with an open and scalable AI platformAdditional feature by Red HatTo be useful in AI models,data needs to be cleansed and normalized,and barriers between databases need to be eliminated.In addition,it

291、s critical that AI solutions provide observability tools for ingesting and processing data at speed and scale.These tools enable real-time monitoring and analysis,ensuring that data flows seamlessly across systems.With a strong data foundation in place,CSPs can use AI to increase automation and oper

292、ational efficiency.AI algorithms and AI-based applications need speed and scale:CSPs need to train,deploy and re-train models quickly;scale the number of models;and manage data across multiple locations.This requires intelligent automation beyond the data center and including the network(for example

293、,being able to automate fault management or network optimization across multiple network domains).3.Integrating AI with existing technologyCSPs must also consider how to integrate their new AI applications and platforms with existing technology in order to minimize risks and increase innovation.This

294、 will allow operators to leverage their internal network-as-a-service transformations to offer NaaS as a capability to customers.These services could range from 5G network slicing to on-demand connectivity to compute-as-a-service.One way to integrate AI with legacy technology is to use APIs that all

295、ow the transfer of information from legacy systems into new ones and vice versa.Turkcell is a good example of a CSP with an AI vision and strategy,enabled by an open platform.The company,which serves nearly 40 million subscribers in Turkey,has been building its AI platform for more than three years,

296、with goals of improving customer experience,increasing operational efficiency and enabling innovation.One area where Turkcell is hoping to capitalize on innovation is with AI chatbots.Today,most commercially available large language models(LLMs)such as OpenAIs ChatGPT and Googles Gemini are trained

297、using English language.But Turkcell serves customers whose primary language is not English.For this reason,the company wanted to use its own data to train chatbots using Turkish language.In the video below,Serkan ztrk,who recently retired as Turkcells CIO,explains the companys approach to AI and the

298、 role for new tools like computer vision,speech and speech-to-text capabilities,and chatbots.He notes that the aim with chatbots is to build“not just a sole chatbot,but a chatbot-as-a-platform that we can also serve and offer outside of Turkcell.”ztrk adds:“In order to support your AI activitiesin t

299、he market,you need to establish a strong platform.We wanted to offer our developers,our coders,a nice environment a comfortable environment for all these coding activities.And the cloud is the best choice to fit this purpose For scalability,for high availability,for efficiency,for efficient resource

300、 usage,this is best with the microservice architecture.So,we started to build our platform based on these necessities.”Turkcell builds an open platform for AI innovationWatch the video to learn more:46inform.tmforum.orgPaving the way to new revenue with an open and scalable AI platformAdditional fea

301、ture by Red HatAnother approach is cloud-native application development,which allows the addition of microservices into legacy applications.In either case,CSPs are able to take an evolutionary path to becoming AI-driven and offering NaaS capabilities beyond connectivity.For example,AI algorithms tha

302、t monitor an enterprise customers network performance could suggest the addition of a location-based service or a better way to route traffic.CSPs could also offer new edge capabilities such as AI-or GPU-as-a-service,where AI compute capacity in edge locations is provided as a service to customers.S

303、umming it upWhile tackling all these challenges,CSPs will need to make organizational transformation a priority.This includes expanding teams skill sets,because selling AI to enterprises and advising them on leveraging AI requires unique expertise.Service providers should focus on differentiating th

304、eir businesses through an open AI platform that will give them the independence they need.Support from a partner like Red Hat can help maximize the value of cloud-native,hybrid-cloud and AI technologies,and set service providers on the right path to become AI native.Learn more about Red Hats AI prod

305、uct portfolio,community projects,partners,training and more.Get in touch to discuss how Red Hat can support you in building your AI strategy and journey.Service providers should focus on differentiating their businesses through an open AI platform that will give them the independence they need.tm fo

306、rum open digital architecture47inform.tmforum.orgThe TM Forum Open Digital Architecture(ODA)provides a migration path from legacy IT systems and processes to modular,cloud-native software orchestrated using AI.ODA comprises tools,code,knowledge and standards(machine-readable assets,not just document

307、s).It is delivering business value for TM Forum members today,accelerating concept-to-cash,eliminating IT&network costs,and enhancing digital customer experience.Developed by TM Forum member organizations through our Collaboration Community and Catalyst proofsof concept,ODA is being used by leading

308、service providers and software companies worldwide.ODA includes:An architecture framework,common language,and design principles Open APIs exposing business services Standardized software components A reference implementation Guides to navigate digital transformation Tools to support the migration fr

309、om legacy architecture to ODA Maturity models and readiness checks to baseline digital capabilities.Goals of the Open Digital ArchitectureThe aim is to transform business agility(accelerating concept-to-cash),enable simpler IT solutions that are easier and cheaper to deploy,integrate and upgrade,and

310、 to establish a standardized software model and market which benefits all parties(service providers,their suppliers and systems integrators).TM Forum Open Digital Architecture-A blueprint for intelligent operationsLearn more about collaborationIf you would like to learn more about the project or how

311、 to get involved in the TM Forum Collaboration Community,please contact George Glass.48inform.tmforum.orgtm forum research reports49inform.tmforum.org50inform.tmforum.orgREPORTDigital Transformation Tracker 7 with automation and Author:Dawn Bushaus,Contributing EditorEditor:Ian Kemp,Managing EditorS

312、ponsored by:TM Forum|June 2023cutting complexity AICSPs take key steps to modernize network inventoryREPORTAuthor:Mark MortensenEditor:Dawn BushausISBN:Sponsored by:June 2023|www.tmforum.orgAuthor:Dr.Mark H.Mortensen,Contributing AnalystEditor:Dawn Bushaus,Contributing Editorsponsored by:Sponsored b

313、y:Author:Editor:ISBN:REPORTtransformingBSS:racing to a flexible,customer-focused futureJune 2023|www.tmforum.orgAuthor:Teresa Cottam,Contributing AnalystEditor:Ian Kemp,Managing Editorsponsored by:Sponsored by:Author:Editor:ISBN:REPORTstandout strategies:how telcos are innovating in a crowded market

314、June 2023|www.tmforum.orgAuthor:Mark Newman,Chief AnalystEditor:Ian Kemp,Managing Editorsponsored by:August 2023|www.tmforum.orgAuthor:Ed Finegold,Contributing AnalystEditor:Ian Kemp,Managing Editorsponsored by:how software-as-a-serviceis reshaping business support systemsReignitingtelecoms growthAu

315、gust 2023|www.tmforum.orgsponsored by:Authors:Sangeet Paul Choudary,Platform Thinking LabsNik Willetts,CEO,TM ForumAnthony Rodrigo,CIO,AxiataDean Ramsay,Principal Analyst,TM Foruma Playbook for CEOsSeptember 2023|www.tmforum.orgAuthor:Mark Newman,Chief AnalystEditor:Dawn Bushaus,Contributing Editor

316、sponsored by:wholesalechanges:rethinking support systemsfor new fiber operatorsSponsored by:Author:Mark Newman Chief Analyst Editor:Dawn Bushaus,Contributing Editor REPORTrethinking support systems for new fiber operatorsSeptember 2023|URL TBCwholesalechanges:Sponsored by:Author:Mark Newman Chief An

317、alyst Editor:Dawn Bushaus,Contributing Editor REPORTrethinking support systems for new fiber operatorsSeptember 2023|URL TBCwholesalechanges:BENCHMARKAuthors:Mark Newman,Chief AnalystDean Ramsay,Principal AnalystEditor:Ian Kemp,Managing EditorTM Forum|September 2023telcorevenuegrowth:time foroperato

318、rs toplace new betsSeptember 2023|www.tmforum.orgAuthor:Patrick Donegan,Principal Analyst,HardenStanceEditor:Dawn Bushaus,Contributing Editor,TM Forumsponsored by:Sponsored by:Author:Patrick Donegan Principal Analyst HardenStanceEditor:Dawn Bushaus,Contributing Editor TM Forum REPORTrisk management

319、moves firmly into the telco spotlightcybersecurity strategies:September 2023|URL TBCrisk management moves firmly into the telco spotlightCybersecurity strategies:REPORTSponsored by:leveling up:Author:Mark Mortensen,Contributing AnalystEditor:Dawn Bushaus,Contributing Editor achieving Level 3 autonom

320、ous networks and beyondAugust 2023September 2023|www.tmforum.orgAuthor:Dr.Mark H Mortensen,Contributing Analyst,TM ForumEditor:Dawn Bushaus,Contributing Editor,TM Forumsponsored by:REPORTSponsored by:leveling up:Author:Mark Mortensen,Contributing AnalystEditor:Dawn Bushaus,Contributing Editor achiev

321、ing Level 3 autonomous networks and beyondAugust 2023REPORTTM Forum|December 2023operators take their first stepsAuthor:Mark Newman,Chief Analyst Editor:Ian Kemp,Managing Editor Sponsored by:generativeAI:REPORTTM Forum|December 2023Author:Ed Finegold,Contributing Analyst Editor:Dawn Bushaus,Contribu

322、ting Editor Sponsored by:BSSfor B2Boperators diverge on the path to cloudREPORTTM Forum|December 2023the sustainable telco:navigating the maze of scope 3 emissionsREPORTa numbers game:February 2024Sponsored by:Author:Richard Webb,Senior AnalystEditor:Ian Kemp,Managing Editor exploiting data to drive

323、 future strategiesFebruary 2024|www.tmforum.orgsponsored by:Author:Richard Webb,Senior AnalystEditor:Ian Kemp,Managing EditorREPORTclosing the loopFebruary 2024Sponsored by:Author:Dean Ramsay,Principal AnalystEditor:Ian Kemp,Managing Editor CSPs aim to automate service orchestration and assuranceFeb

324、ruary 2024|www.tmforum.orgsponsored by:Author:Dean Ramsay,Practice Lead,TM ForumEditor:Ian Kemp,Managing EditorREPORTclosing the loopFebruary 2024Sponsored by:Author:Dean Ramsay,Principal AnalystEditor:Ian Kemp,Managing Editor CSPs aim to automate service orchestration and assuranceREPORTAuthor:Dawn

325、 Bushaus,Contributing AnalystEditor:Ian Kemp,Managing EditorSponsored by:inside thetelco talentrevolutionfinding skills for the future:TM Forum|May 2024April 2024|www.tmforum.orgAuthor:Charlotte Patrick,Contributing AnalystEditor:Dawn Bushaus,Contributing Editorfocus onsponsored by:driving intellige

326、ncein network lifecycleautomation:Asia-Pacificinnovation:the future for telecoms R&DREPORTin search ofTM Forum|July 2024Sponsored by:Author:Ed Finegold,Contributing AnalystEditor:Ian Kemp,Managing Editor July 2024|www.tmforum.orgAuthor:Ed Finegold,Contributing AnalystEditor:Ian Kemp,Managing EditorR

327、EPORTAuthor:Dawn Bushaus,Contributing AnalystEditor:Ian Kemp,Managing EditorSponsored by:cloud adoption acceleratesTM Forum|June 2024Digital Transformation Tracker 8:September 2024|www.tmforum.orginnovating in the Americas:key driversfor growthAuthor:Charlotte Patrick,Contributing AnalystEditor:Ian

328、Kemp,Managing Editorexploring the September 2024|www.inform.tmforum.orgSponsored by:Author:Charlotte Patrick,Contributing AnalystEditor:Ian Kemp,Managing Editor key drivers for growthinnovating in the Americas:REGIONAL REPORTData&AI Innovationexploring theREGIONAL REPORTREPORTAuthor:Teresa Cottam,Co

329、ntributing AnalystEditor:Ian Kemp,Managing EditorSponsored by:TM Forum|September 2024putting people at the heart of customer experiencehumanizing AI:Sponsored by:Telco revenue growth:October 2024|https:/inform.tmforum.orgAuthors:Mark Newman,Chief AnalystRichard Webb,Senior AnalystEditor:Ian Kemp,Man

330、aging Editor Is B2B a beacon of light?BENCHMARK REPORTNovember 2024|www.tmforum.orginnovating in Asia:key driversfor growthAuthors:Mark Newman,Chief Analyst Charlotte Patrick,Contributing AnalystEditor:Ian Kemp,Managing Editorsponsored by:exploring the September 2024|www.inform.tmforum.orgSponsored

331、by:Author:Charlotte Patrick,Contributing AnalystEditor:Ian Kemp,Managing Editor key drivers for growthinnovating in the Americas:REGIONAL REPORTData&AI Innovationexploring theREGIONAL REPORTexploring the November 2024|www.inform.tmforum.orgSponsored by:Author:Charlotte Patrick,Contributing AnalystEd

332、itor:Ian Kemp,Managing Editor key drivers for growthinnovating in Asia:REGIONAL REPORTexploring the November 2024|www.inform.tmforum.orgSponsored by:Author:Charlotte Patrick,Contributing AnalystEditor:Ian Kemp,Managing Editor key drivers for growthinnovating in Asia:REGIONAL REPORTOctober 2024|www.t

333、mforum.orgMeetingexpandedexpectationsAuthor:Ed Finegold,Contributing AnalystEditors:Ian Kemp,Managing EditorDawn Bushaus,Contributing Editorsponsored by:for networkchargingDecember 2024|www.tmforum.orgtowardsautonomyAuthor:Richard Webb,Senior AnalystEditor:Ian Kemp,Managing Editorsponsored by:in mobile privatenetworksin mobile private networksDecember 2024|www.inform.tmforum.orgSponsored by:Author