2024探索自智網(wǎng)絡(luò)的最佳實(shí)踐研究報告：完全自智網(wǎng)絡(luò)發(fā)展路徑中的運(yùn)營商定位（英文）

tmforumDecember2024|tmforumBENCHMARKREPORTAutonomousnetworkoperationsAuthors:RichardWebb,SeniorAnalystSponsoredby:DawnBushaus,ContributingAnalystSponsoredby:NO<INO<IAIanKemp,ManagingEditorcontents03aboutoursponsorssection2:surveyresults–whereareCSPsontheANjourney?04thebigpicturesection3:creatingablueprintforLevel4autonomousnetworks07section1:autonomousnetworks25section4:ANbestpracticeinaction32section5:award-winningCatalysts35section6:collaborationonstandardsiskeytoANinteroperability38section7:keyfindingsandrecommendations40additionalfeatures&resources22BENCHMARKBENCHMARKaboutourAtNokia,wecreatetechnologythathelpstheworldacttogether.Weputtheworld’speople,machinesanddevicesinsynctocreateamoresustainable,productiveandaccessiblefuture.WeareaB2Btechnologyinnovationleaderpioneeringnetworksthatsense,thinkZTEZTECorporationisaleadingglobalprovideroftelecommunicationssolutionsandservices,deliveringinnovativetechnologiesthatpowernetworks,enterprises,andconsumersworldwide.Establishedin1985,ZTEoperatesinover160countries,specializingincomprehensiveend-to-endsolutionsacrossfixed,mobile,IT,AIandenterprisenetworks.Withexpertisein5G,AutonomousNetworks(AN),cloudcomputing,AI,andIoT,ZTEempowerscustomerstoachievenetworkintelligence,operationalefficiency,andsustainability.Renownedforitscommitmenttoresearchanddevelopment,ZTEdrivescutting-edgeadvancementsthatenableseamlessconnectivityanddigitaltransformation,buildingasmarter,moreconnectedfutureforitsglobalpartnersandcommunities.Tolearnmoreaboutthebenefitsofautonomousnetworksandhowoursponsorsarehelpingtofacilitatetheirdevelopment,watchthesevideos:NokiaZTE3BENCHMARKBENCHMARKthebigpictureInthepastcoupleofyearswehavepublishedreportsthathavesetoutclearlythechallengesforcommunicationsserviceproviders(CSPs)inthefaceofincreasingnetworkcomplexity,limitedreturnoninvestmentin5Gandgrowingexpectationsforimprovedcustomerexperience.OurannualtelcorevenuegrowthBenchmarkreportshavehighlightedhowdifficultitisforCSPstoestablishnewrevenuestreamsanddriveoperationalefficiencyinordertocompetewithmoreagilestart-upsandhyperscalecloudserviceproviders.Morerecently,theemergenceofAIhasprovenhighlydisruptive,butalsorepresentssignificantopportunitiesforoperators.ThisreportdrawsonsomeofTourresearchthisyeartoexplorehowAIand,moreexplicitly,automationcanaddressthesechallenges.ItlooksatwhatCSPsaredoingintermsofnetworkautomationbothwithinTMForum’smembercommunityandmorewidely,andattemptstoidentifysomeofthebestpracticesofoperatorsstrivingtogobeyondsimpleautomationtoimplementautonomousnetworks(AN).Aspartofthisnewresearchweconductedasurveywhichdrewresponsesfrom111individualsat82companies(seesection2).AmongtheprimarydriversforthoserespondentstoimplementANareimprovingcustomerexperience/satisfaction;simplifyingoperationsandmaintenanceandimprovingpersonnelefficiency;andincreasingresourceefficiencyandutilizationofnetworkandIT.Amonglargeoperators–thosewithmorethan100millioncustomers–100%saidsupportingagileservicedeliveryandacceleratingtimetomarket/monetizationwasthemostimportantdriver.Buttherearemanyotherreasons.Autonomousnetworksarenotsimplyatechnicalevolution;theyrepresentatransformativechangeinhownetworksoperate,adaptandserveCSPsandtheircustomers.Whereasautomationreliesonpredefinedrules,networkautonomyinvolvesintelligentsystemsmakingindependentdecisions–withoutinputfromhumans.Inbrief,muchofthevalueforCSPsliesintheabilityofANtoaddressthreecriticalareas:Operationalefficiency.ANoptimizesresourcemanagementthroughautonomousprocesses,managingnetworkcomplexityandenablingCSPstosignificantlyreducemanualinterventionsandoperationalcosts.Customerexperiencetransformation.Customersdemandreliable,seamlessconnectivityandpersonalizedservices.AN,basedonanintent-drivenarchitecture,enablesCSPstomeettheseexpectationsbyproactivelymanagingnetworkqualityanddeliveringservicelevelguarantees.Revenueandinnovationpotential.ANunlocksnewbusinessmodels,suchasnetwork-as-a-service(NaaS),enablingCSPstomonetizetheirnetworksbeyondtraditionalconnectivity.4thebigpictureMovinguptheANlevelsTMForumisattheforefrontofindustry-widecollaborationtodrivethedeploymentandevolutionofANtechnology,processesandbusinessmodels.AttheheartofthatworkistheAutonomousNetworksProject.Ultimately,CSPsneedtobeabletoscaleorchestrationofzero-touch,zero-waitandzero-troubleservicesendtoendacrossdifferentnetworkdomains(forexample,radioaccess,fixedaccess,core,IP,opticaltransportanddatacenter/cloudnetworks).Thisrequiresautomationoftheentireservicelifecycle,fromorderingtofulfillment,activation,orchestration,management,assurance,optimizationandbilling.InconjunctionwithitscommunityofCSPandsolutionprovidermembers,theForumhasdevelopedasix-leveltaxonomyforCSPstomeasuretheirprogressinimplementingautonomousnetworks.EachlevelhasasetofcharacteristicsdescribingtheevolutionarystageoftheCSP’sjourneyfrommanualtofullyautonomousoperations(seegraphic).AccordingtoTMForum’sdefinition,Level4autonomousnetworksrepresentamajorshiftfromtraditionalhuman-definedautomationprocessestotrueautonomousdecision-making.ThisisasignificantleapfromLevel3,wheremachinesassisthumansindecision-makingbutstillrelyonhumanoversight.Today,mostoperatorsarebetweenLevels2and3.Butuntilrecentlyit’sbeendifficulttogaugeCSPs’progressdeployingANbecausetherewasnotastandardizedwaytoevaluateandcompareimplementations.Thatchangedin2024withfurtherstandardizationofTMForum’smethodologyandtoolsforevaluatingANlevels,aswe’llseeinsections2and3ofthisreport.Fullyautonomousnetwork:Thesystemhasclosed-loopautomationFullyautonomousnetwork:Thesystemhasclosed-loopautomationcapabilitiesacrossmultipleservices,multipledomains(includingpartners’domains)andtheentirelifecycleviacognitiveself-adaptation.Highlyautonomousnetwork:Inamorecomplicatedcross-domainenvironment,thesystemenablesdecision-makingbasedonpredictiveanalysisoractiveclosed-loopmanagementofservice-drivenandcustomerexperience-drivennetworksviaAImodelingandcontinuouslearning.Conditionalautonomousnetwork:Thesystemsensesreal-timeenvironmentalchangesandincertainnetworkdomainswilloptimizeandadjustitselftotheexternalenvironmenttoenable,closed-loopmanagementviadynamicallyprogrammablepolicies.Partialautonomousnetwork:Thesystemenablesclosed-loopoperationsandmaintenanceforspeci?cunitsundercertainexternalenvironmentsviastaticallycon?guredrules.Assistedoperationsandmaintenance:Thesystemexecutesaspeci?c,repetitivesubtaskbasedonpre-con?guration,whichcanberecordedonlineandtraced,inordertoincreaseexecutionefficiency.Manualoperationsandmaintenance:Thesystemdeliversassistedmonitoringcapabilities,butalldynamictasksmustbeexecutedmanually.54321055thebigpictureIndustrycollaborationonANMorethan60CSPsworldwidehavesignedTMForum’sANManifesto,pledgingtoworktowardsLevel4autonomy,anddozensareparticipatingintheForum’sANlevelevaluation.EvenmorehaveparticipatedinANCatalystproofofconcepts.Morewidely,acrossthetelecomsindustry,morethan120ANstandardsandresearchprojectshavebeeninitiated,withmultiplestandardsdevelopmentorganizationsandopen-sourcegroupsworkingtodefineanddeployANs(moreonthisinsection6).WithinTMForum’sAutonomousNetworksProjectmanyCSPshaveoutlinedtheirANvisionsandstrategies,and24havecarriedoutinternalANlevelassessments.AhandfulofCSPshavepubliclystatedtheyareadoptingANLevel4“high-valuescenarios”,aimingtoreachLevel4foroperationalflowsinsomedomainsfrom2025to2027.AtLevel4,thenetworkcanself-manage,self-optimizeandhandlecomplextasks.Toachievethislevelofautonomy,CSPsneedtotranslatetheircustomers’businessgoals,or“intents”,intonetworkrequirementsthatcanbefulfilledwithoutcustomersneedingtounderstandtheunderlyinginfrastructure.DaveMilham,TMForum’sChiefArchitect,describesthistransition:“MovingtoANLevel4’sintent-basedinterfacesisamovefromprescriptivesupervisionautomationtodeclarative,delegatedautonomy,inwhichAIdoesthereasoning.”Transitioningfromautomationtoautonomyisasignificantbusinessandtechnologychallenge.InNovember2024,TMForumpublishedablueprintframeworkforachievingANLevel4,whichemphasizesAIbeingembeddedintoalldomaindesign,operationandprocesses,suchasclosed-looporchestration,toenableintent-basedservices.Section3providesanoverviewofthatblueprint.“ItisvitalthatANleveragesintent,sothatinsteadofsellingtechnologyCSPsaresellingconnectivitythatcanbetailoredtobusinessoutcomes,”saidTMForumCTOGeorgeGlassinhiskeynoteaddressatInnovateAsia24.TheimportanceofintentinAI-drivennetworksandservicesisdiscussedfurtherinthenextsection.Thisinter-dependencybetweenAIandANisattheheartofanon-goingindustrydialogaround‘AIfornetworks,versusnetworksforAI’.AIiscriticalfortheevolutionofAN,butANisalsocriticaltosupportfutureAI-drivendemandsonconnectivity–inotherwords,ensuringthatnetworksareAI-ready.CSPsmustadoptAIfornetworks,shiftingfromstatic,rules-basednetworkautomationtoAI-enabled,dynamicorchestrationacrossmultiplenetworkdomains–from5GRANtoIPtransportandopticalcore.WelookattheroleofAIinautonomousnetworksinthenextsection,aswellasthebusinesscaseforAN.Readthisreporttounderstand:ThebusinessgoalsandvisionsdrivingANadoptionWhichareasCSPsareprioritizingforANimplementationTherelationshipbetweenAIandintent-basedautomationTheblueprintforachievingANLevel4HowANbestpracticeisbeingdefinedinreal-worldnetworkdeployments.TolearnmoreabouthowCSPsaretransitioningtoLevel4autonomousnetworksreadTMForum’sJourneyGuide:66BENCHMARKBENCHMARKsection1:autonomousnetworksFormostCSPs,adoptionofAIcomeswithadualfocus:usingAItooptimizetheirnetworks(aspartoftheirtransformationtowardsAN);andensuringthattheirnetworkscansupportthedynamicrequirementsofemerging–andpotentiallyrevenue-generating–AI-poweredapplications.noursurveyforthisreport,weaskedaboutthechallengesCSPsfacewhenimplementingAN.Thetopthreechallengesarecomplexintegrationchallengesacrossdomains,lackofbudget,andalackofaclearpathtoLevel4AN.AntonReynaldoBonifacio,EVP,ChiefInformationSecurityOfficerandChiefAIOfficer,GlobeTelecom,speakingatInnovateAsia24,emphasizedtheimportanceofestablishinganAI-firstculturetotacklesomeoftheinherentchallengesandenhancegrowthandsecurityinautonomousnetworks.InGlobe’scase,thisisbasedonademocratized,bottom-upapproachthatismoretacticalthanstrategic.ComplexintegrationchallengesacrossdomainsLackofComplexintegrationchallengesacrossdomainsLackofbudgetTheend-to-endarchitectureisnotde?ned,andthereisnoclearpathtoLevel4Keytechnologiesrequired(suchasAI)arenotmatureWedonothavetherightskillsLevelevaluationisdi?cult/welackthemethodologyortoolstomakeacredibleevaluationLackofclearbusinessobjectivesandvalueindicatorsOurdatastrategydoesnotadequatelysupportANimplementationANstandardsarenotmatureenoughforimplementationLackofusecasesthatcanbeappliedatscaleLackoftop-levelstrategyandvision53%43%4%43%48%43%48%9%34%34%52%14%29%29%60%10%26%26%58%16%24%58%18%24%58%18%23%23%51%26%23%53%23%53%25%22%22%60%18%19%19%57%25%18%18%37%45%SeriouschallengeSomewhatofachallengeNotachallengeTMForum,202477Improvingcustomerexperience/satisfactionSimplifyingoperationsandmaintenanceandimprovingpersonnele?ciencyReducingthecostsofoperationsandmaintenanceIncreasingresourcee?ciencyandutilization(networkandIT)Serviceoptimizationandensuringqualityofservice/qualityofexperienceSupportingagileservicesdeliveryandacceleratingtimetomarket/monetizationFaultmanagementusingautomatedclosed-loopoperationsGreenenergysavingtomeetsustainabilitygoalsEnablingexposureofnetwork-as-a-service(NaaS)capabilitiesEnabling/enhancinginteractionwithpartners85%84%15%84%15%83%17%83%17%78%78%20%70%28%70%28%53%41%6%53%41%6%50%38%50%38%12%42%42%46%12%VeryimportantSomewhatimportantNotimportantTMForum,2024“WeequipemployeeswithAIcapabilitiestousewheretheyfeelitisappropriateratherthandictateusecasestofocuson,”saidBonifacio.“Not‘chasingROI’butconsideringAIintermsofemployee,team,businessorcustomerimpact.Itisimportanttohavea‘playground’approachtoallowforsomeopennesstoexplore,butwithinsafetynetparameters,withnorisktocustomers.”TheabilitytoscaleusecasesforANimplementationiscriticaltothebusinesscase.CSPshaveorganizedtheirIToperationsaroundlinearsoftwareprocessesmanagedbyseparateteams(fulfillment,serviceactivationandassurance,forexample).Thesesilosexistineverydomain,butCSPsneedtobeabletoscaleorchestrationofzero-touch,zero-waitandzero-troubleservices,end-to-endacrossnetworkdomains.Thisrequiresautomationoftheentireservicelifecycle,fromorderingto88fulfillment,activation,orchestration,management,assurance,optimizationandbilling.“Wehaveanend-to-endserviceoperationmanagementlayeronthetopofthefivedifferenttechnicaldomains,andwearespecifyingforeachyearthetechnicalmaturitylevelrequirements,”saidDr.LingliDeng,DirectorandResearcheratChinaMobile,inaninterviewforTMForumInform.ShesaidChinaMobileisusingitsend-to-endserviceoperationlayerasthecatalysttodriveoveralldevelopmentoftechnicalANcapabilities.Despitetheimplementationchallenges,ANsarecrucialforCSPstorealizethefullbusinesspotentialoftheirinfrastructureinvestments.Inoursurvey,weaskedrespondentsabouttheirANimplementationdrivers.Perhapsunsurprisingly,thetop-ratedANdriverrelatestoenhancingcustomerexperience.AutonomousnetworkscanhelpCSPsunderstandsubscriberbehaviorandtakeproactivemeasurestoimprovecustomersatisfaction.Forinstance,AI-poweredanalysisenablesCSPstoassessnetworkperformance,suchaslatencyatthecustomerlevel,predictfailuresandimprovenetworkreliability.InakeynoteaddressatInnovateAsia24,ShantiJuanitaJohari,CCO,ConsumerStrategyandBusiness,TelekomMalaysia,spokeaboutdrivingaserviceculturethroughAIandAN.“We’vemadeacustomerpromisetorestorefixedbroadbandfaultswithin24hoursoryourmoneyback–quiteaboldservicedifferentiator.‘Owning’thatpromiseisimportant:todriveacustomer-centricculturewithinTelekomMalaysiaandgetemployeesthinkingaboutthecustomerimpactofautomationactivities.”Intent-based,closed-loopmanagemeTMForum,2024Headdedthatdoingsohaschangedtheinternalperspective,from“afearoflosingmoney,topositiveconversationaboutwhatwestandtogain.EngineeringteamsaremotivatedbytougherKPIs,andwearenowimplementingthistypeofpromiseinotherareasofthebusiness.”Thenexttwoleadingdriversforoursurveyrespondentsareinterconnected:operationalsimplicityandcostefficiency.ANsenableCSPstooptimizeoperationsandmaintenance,reduceoperationalexpenditure,andimproveprofitability–allkeybenefitswhilerevenuefromtraditionaltelecomsservicesremainsstatic.Withautonomousnetworking,CSPsdonothaveto99useresourcestoconductroutinetaskssuchasprovisioningandfaultdetection,loweringcostsandtimetorectifyproblems.Intent-basednetworksIntentisthedrivingforceofAN,communicatingrequirements,constraintsandpreferencestoanautonomoussystem,enablingCSPstosetthegoalsandparametersforAImodelswithoutprescribingspecificactions.Thisallowsthenetworktoadaptmorefluidly,andenablesCSPstoturntheircustomers’businessgoals,orintents,intoinstructionsforthenetworkthatcanbefulfilledwithoutcustomersneedingtounderstandtheunderlyinginfrastructure–abstractingitscomplexityandcommunicatingwithcustomersusingacommonlanguageorontologythatcanalsobemachinereadable.ThegraphicshowstheinteractionofintentsacrossaCSP’soperationallayers.Intent-basedmanagementisbakedintotheANjourneyatleastasearlyasLevel3ofthesix-levelmodeldescribedearlier(conditionalautonomousnetworks)andbecomesessentialatlevels4and5–highlyandfullyautonomousnetworks,respectively.Tobeviable,intentsmustbegroundedindata:observed,measuredandcontrolledintentswithwhichtodrivetheautonomoussystem,andverifiableunderdynamicallychangeableconditions.Andtomanagetheseprocessesthedatamustbeaccurateandeasilyaccessible.Toautomatethefulllifecycleofservices,intentsarecombinedwithclosed-loopmanagement.Closingtheloopmeanscollecting/analyzingdatatofigureouthownetworkscanbeoptimizedandimplementingthosechangesinanautomatedway.Intentcanbeexpressedatvariousoperationallayers(businessintent,serviceintentandresourceintent,forexample).InTMForum’sANReferenceArchitecture(IG1251),intentsguidetheclosedloopcontrolprocess,whichisacontinuousprocessofmonitoring,analyzingandadjustingperformancetoimprovethequalityofexperienceforendusersandautomatethefulllifecycleofservices.“Intent…isourdream,ouraspiration,butitdependsonthematurityoftheinfrastructureitself,”saidIndraMardiatna,CTOofIndonesianoperatorTelkomsel,inaninterviewwithTMForumInform.“Maybesomeusecaseswecangettothatlevelby2025,butinsomeverycomplexcases[we]requiremorepower…andmoredata,andweneedtime.”UsingintentforslicingNetworkslicingwillbeoneofthefirstusecasesforintent-basedautomationtoenablethecreationofmultiplevirtual(orlogical)networkswithinasharedphysicalinfrastructure.Asnetworkslicesproliferate,CSPswillneedtomanageandorchestratenetworkresources.AIcanhelpmanagethiscomplexitybypredictingtrafficflowsandrequirements,networkissuesandoutages,andsendinginstructionstodifferentnetworkfunctionsaboutsolutions.“Networkslicingbringsmanagement,operationsandmaintenancecomplexity,”saysYessieYosetya,DirectorandChiefEnterprise&CorporateAffairsOfficeratXLAxiata.“Multi-dimensionalmanagementofhardware,resources,slicedefinitionandapplicationsarerequired.OnlyANscanscaleuptothedemandsofprovisioningandmaintainingthesliceinrealtime.”CSPswillneedintent-drivenorchestrationtosupportdynamicslicing.Whilstinitialdeploymentsofslicesinthenetworkcanbedonewithoutfullautomation,torepeatthisatscale,andhandledifferentcustomerrequirements,itwillneedtobeintent-driven.TheroleofGenAIinANsGenerativeAI(GenAI)willplayanincreasinglyimportantroleinAN,workingtogetherwithotherAIandmachinelearningtechnologies.Machinelearningcanbeusedtoidentifypatternsindata,predictresultsbasedonhistoricalinformationandmakedecisions,whileGenAIiswellsuitedtotasksrequiringhighcreativityandinnovation,suchasintelligentQ&Aforroutineoperationsandmaintenancetasksandcomplexnetworkconfiguration.OursurveyaskedrespondentswhetherGenAIisbeingintroducedintonetworkoperations.Intotal,64%ofrespondentssaidtheyhadeitheralreadyimplementeditorwereplanningtodoso.ButthatstillleavesonethirdwhothinkthatGenAIiseithernotmatureenoughforuseinnetworkoperationsyetorshouldnotbeusedinmission-criticalnetworkfunctions(seepiechartabove).Combined,“traditional”AIandGenAIwillbeabletoidentifyconstantlychangingnetworkrequirementsandgenerateoptimalnetworkconfigurations.TheycanalsohelpCSPsimprovenetworkperformanceandreliabilitybygeneratingcodeandCSPs'introductionofGenAIintonetworkoperations8%8%27%47%WehaveintroducedGenAIinnetworkoperationsWeplantointroduceGenAIGenerativeAIisnotmatureenoughforuseinnetworkoperationsyetThetrustworthinessofGenAIistoouncertainforuseinmission-criticalnetworkfunctionsTMForum,2024analyzingdatatodetectnetworkproblems.Forexample,anLLMcangeneratenetworktopologydiagramsorbuildadigitaltwinbasedontextdescriptionsandlearnpatternsfromhistoricaldatatoidentifyanomaliesandtherootcauseofproblems.ChinaTelecom,forexample,isusingGenAItobuildadigitaltwinofitsnetworktoenablerootcauseanalysis.WeaskedCSPshowtheyarealreadyusingGenAIinthenetwork.Themajority(74%)saidtheyareusingitfornetworkmaintenance,suchasautomatedtroubleshootingandriskprediction.Ahighproportionarealsousingitfornetworkoperations,suchasintelligentcustomerservice,andnetworkoptimization,suchasriskidentificationandautomatedverificationofoptimization(seegraphiconthenextpage).We’lllookatsomedifferencesinhowsmallandlargeoperatorsviewGenAIinthenextsection.Indeed,GenAIhasmultipleotherpotentialapplicationsfornetworkoperations:Decision-making.Byinterpretingcomplex,unstructureddatasuchasfieldengineerreports,GenAIcanrefineAImodelsforbetterresourceallocationandfaultdiagnosis.Naturallanguageinteraction.GenAIcanprovideintuitiveinterfacesforCSPs,allowinginteractionwithsystemsusingnaturallanguagequeriessuchas“HowcanIimproveservicequalityinthisarea?”Serviceinnovation.GenAIcananalyzemarkettrendsandsuggestnewserviceofferingsornetworkconfigurationstailoredtoemergingcustomerneeds.Securityandfrauddetection.GenAIcanbeusedtoanalyzeterabytesofnetworkdatatodetectanomalies,enhancesecurityandimprovefrauddetection.Softwarecoding.GenAImodelscanbetrainedonlargecoderepositoriestogivecompletionpromptsasdeveloperswritecode,ortogeneratenewcode.Real-timelanguagetranslation.Itcanalsotranslatelanguagesinrealtime,suchasdocumentation,HRquestions,summarizingmeetingsandsoon.InanarticleforTMForumInform,UsmanJavaid,ChiefProductsandMarketingOfficer,OrangeBusinessandBrunoZerbib,thecompany’sGroupCTIO,arguethatnetworksneedtochangetoaccommodateaneweraofGenAIandquantumcomputing.“Networkserviceprovidersmustlaythefoundations,standards,androadmapforanetworkforAIthatisdistributed,scalable,secure,andenergyefficient,”theyHowareCSPsusingGenAIinthenetwork?●·62%·●Optimization46%OptimizationDeploymentOperations (e.g.,intelligentcustomerservice)Planning65%DeploymentOperations (e.g.,intelligentcustomerservice)Planning65%30%IntelligenttrainingservicesceIntelligenttrainingservices2024TMForum,2024membersarecollaboratingonanindustry-specificdatareferencearchitecture,encirclingbothemergingAI-enabledbusinessmodelsandsupportingnetworks.”DataiskeyInTMForum’sModernDataArchitectureProjectmembersarecollaboratingonatelecoms-specificdatareferencearchitecture,incorporatingemergingAI-enabledbusinessmodels.TheprojectcomesfromtheneedtomodernizedataarchitecturestosupporttherapidevolutionofAIandhelpdefinecutting-edgeAI-enabledtelecomsoperations.Theultimateaimistorealizezero-touchefficienciesforafullyAI-enableddataarchitecture.LackofanadequatedatastrategywasratedasaseriouschallengeorsomewhatofachallengeforANimplementationby76%ofrespondentstooursurvey.Indeed,establishingarobustdataculturewithintheCSPiscritical.“Astrongdatafoundationisfundamentaltointelligentoperations.Itunderpinsourwholeanalyticsandautomationstrategy,”saysOlivierSimon,SVPSmartNetwork&Data,OrangeIn