IBM：2025年亞太地區(qū)人工智能展望

APACAIOutlook2025

THENEXTFRONTIEROFAIRACE:

SCALINGAIFORIMPACT

NOVEMBER2024

AUTHOREDBY

SASHMUKHERJEE

VPIndustryInsights,Ecosystm

SPONSOREDBY

2

CASESTUDIES

INTRODUCTION

THENEXTWAVEOFAI

SCALINGFORSUCCESS

Introduction

ImagineafuturewhereAIdrivesrealbusinessimpact,notjustbuzz.Whileitspotentialisevident,manyorganisationsstillstruggletoharnessiteffectively.

In2024,experimentationandtrend-chasingdefinedtheAIlandscape.In2025,thefocuswillshifttodeliveringtangible

valuethroughrobustinfrastructure,efficientoperations,andskilledtalent.Successwillhingeonastrategicapproach:clearoutcomes,strongdatamanagement,andgovernance.

ThiswhitepaperexamineskeyAItrendsandthechallengesorganisationsmustaddresstounlockAI'stransformativepotential.Insightsfrom17APACorganisationsprovideablueprintforacceleratingAIinitiativeswhilemanagingriskseffectively.

ECOSYSTMAPACAIOUTLOOK2025|IBM

TheNextWave

ofAI:Whatto

Expectin2025?

3

INTRODUCTION

THENEXTWAVEOFAI

SCALINGFORSUCCESS

CASESTUDIES

LargeenterpriseshavebeenactivelyexperimentingwithAIin2024,particularlyinthewakeofGenAI’srapidadvancement.Initialenthusiasm,fuelledbybusinessmandatesandreadilyavailabletechnology,ledtoaflurryofAIprojects.

However,astheyearprogressed,amorenuancedunderstandingofAI'spotentialandchallenges

emerged.Whilethefocusremainsonidentifyingusecasestoenhanceemployeeproductivityand

customerexperience,organisationsarenowprioritisingfoundationalelementslikedatagovernance,

dataquality,andskilledtalent.Moreimportantly,theemphasishasshiftedtomaximisingROIfromAI

investments,giventhesignificantresourcerequirements.Theneedforopen-sourceAIandtheabilitytointegrateAIplatformsfromanytechproviderhasalsobecomeincreasinglyimportant,mostlytoavoid

vendorlock-ins.

AIWake-upCall:RethinkYourStrategy

Wehaveenoughusecases-weneedtoprioritisethembasedondataavailability,securityandprivacyconsiderations,andROI.

BANKINGCDO,SINGAPORE

OurAlinitiativeshavesloweddown.Weareextremelyconcernedaboutdatagovernancestartingfirstwithprivacyandsecurity.

INSURANCECIO,AUSTRALIA

AlhastakenusbacktotheearlyClouddays-wearestrugglingwithprotectingourorganisationfromShadowITandBYOAI.

RETAILCTO,INDIA

ECOSYSTMAPACAIOUTLOOK2025|IBM

THENEXTWAVEOFAI

SCALINGFORSUCCESS

Whileorganisationshavestruggledtorealisetheirexpectedbenefits,technologyvendorscontinuetoinnovate,makingAImoreaccessibletoenterprises.

Examiningthekeytrends–bothorganisationalandtechnological–expectedtoimpacttheAIlandscapein2025canguideenterprisesonhowtocalibrateorstarttheirAIjourneys,andmoreimportantly,wheretobegin.

Herearethe5keytrendsthatwillimpacttheAIlandscapein2025:

StrategicAI

MaximisingImpact

RightsizingAI

Targeted,Open-

SourceAIModels

forEfficiency

UnifiedAI

Ensuring

Managementand

Governance

AgenticAI

Empowering

IntelligentSystems

Beyond

Productivity

TheHuman-Centric

FutureofAI

4

INTRODUCTION

CASESTUDIES

ECOSYSTMAPACAIOUTLOOK2025|IBM

THENEXTWAVEOFAI

SCALINGFORSUCCESS

#1StrategicAI:MaximisingImpact

OrganisationswilladoptamorestrategicapproachtoAI,prioritisingprojectsbasedonfeasibilityandbusinessimpact.

AI'slong-termbenefitsandhighupfrontcostschallengetraditionalROImetrics,oftenleading

businessleaderstopushforearlyresultswithoutgraspingthecomplexities.Toaddressthis,

techleadershavetraditionallyfocusedonquick-winusecasestobuildtrustandinternalbuy-in.Asorganisations’AIjourneysmature,theywillaimtobalanceshort-termwinswithlong-termAIstrategies.ThefocusofAIinvestmentsisshiftingbeyondemployeeproductivityandcustomer

experience,towardsbroaderstrategicgoalssuchasinnovationandimpactoncompanyfinancials.

Nearly60%ofAsiaPacific

organisationsanticipaterealising

thebenefitsoftheirAIinvestmentswithin2-5years.Only11%expect

immediatereturnswithinthenexttwoyears.

SOURCE:ECOSYSTM,2024

BEYONDTHEIMMEDIATE:LONG-TERMBENEFITSOFAI

36%

32%

26%

Innovationofservice/product/businessmodel

21%21%

18%

Increasedrevenue

12%11%12%

Costsavings

12%

7%

9%

Increasedemployeeproductivity

3%3%3%

Improvedcustomerexperience

5

INTRODUCTION

CASESTUDIES

ANZASEANIndia

ECOSYSTMAPACAIOUTLOOK2025|IBM

Q:WhatbenefitsdoesyourorganisationexpectfromAIinthenext2years?N=518;Source:Ecosystm,2024

Wewillseeashiftfromlow-risk,non-coreusecasesto

deployingGenAIincorebusinessfunctionsforcompetitiveadvantageandimprovedROI.

TechanddataleaderswilladoptcomprehensiveAIevaluationframeworks,assessingfinancialmetricsalongsidebroaderimpactslikejobroles

anddatagovernance.Selectingtherightusecaseinvolvesatwo-step

process:prioritisingwithstructuredassessmentsandevaluatingtechnicalfeasibility.Thiswillincludeexaminingdatausability,infrastructure,digitalinvestments,processreadiness,andresourceneeds.

TraditionalROImetricsstrugglewithAI’slong-term,intangiblebenefitsandhighupfrontcosts.WhilePoCsvalidatefeasibility,theyoftenmissscaling

complexitiesandtruecosts.Toaddressthis,organisationswillembrace

morenuancedevaluationapproaches,balancingtangibleandintangiblebenefits.Aholisticcostingstrategyinvolvingbusiness,technology,data,andfinanceteamswillbeacriticalaspecttoaccountforinfrastructure,

hardware,software,andpersonnelexpenses,acrosstheprojectlifecycle.

6

INTRODUCTION

THENEXTWAVEOFAI

SCALINGFORSUCCESS

CASESTUDIES

CASESTUDIES

StrategicAdoptionofAI

StarUnionDai-ichiLifeInsurance(SUDLife)focusedon

tacklingaspecific,highimpactusecase–thechallenge

ofoutperforminglarge-capportfoliosinIndia’scompetitive

capitalmarkets.ByleveragingGenAI,thecollaborationaimstodelivercriticaldata-driveninsightsforanewinvestment

product.Outcomesincludeenhanceddataanalysisfor

extractingvaluableinsights,improveddecision-making

throughdata-driventoolsforfundmanagers,andadherence

toresponsibleAIpractices.AIhasbecomeanessentialtool

forfundmanagerstonavigatevastdatavolumes,transformingitfroma"nice-to-have"toa"must-have."

StarHubhasstrategicallyintegratedAItoenhancecustomer

experience,streamlineprocesses,anddriveinnovation.TheirCloudInfinity,theworld’sfirstmetropolitanhybridmulti-cloudarchitecture,usesAIforautomatedresourcemanagement,

allowingenterprisestoefficientlyscaleresourcesandoptimiseapplicationsanddata.TheyarealsocollaboratingwithamajorretailoperatortodeployaSmartRetailsolutionthatcombines

GenAIandbusinessintelligence,generatingactionableinsightsfromcustomerdata.

ECOSYSTMAPACAIOUTLOOK2025|IBM

7

INTRODUCTION

CASESTUDIES

THENEXTWAVEOFAI

SCALINGFORSUCCESS

#2RightsizingAI:Targeted,Open-sourceAImodelsforEfficiency

Smaller,open-source,specialisedmodelswillgaintraction,offeringabalancebetweenperformance,resourceefficiency,andflexibility.

OrganisationsinAsiaPacificwillincreasinglyleverageopen-sourceAImodelstodriveinnovationandefficiency.Thiswillbeagame-changer,offeringcost-effectiveness,seamlessintegrations,andtheflexibilitytousecustommodelsorleveragevendor-specificcapabilities.

Whilelargelanguagemodels(LLMs)havecapturedtheimagination,smaller,specialisedmodelstailoredtospecifictasksordomainsofferacompellingalternative.Thesemodelsoftendelivercomparableperformancewhilerequiringlesscomputationalpower,makingthemidealfororganisationsaimingtotrainmodelsusingproprietarydata.Additionally,theyaremoreenergy-efficient,aligningwithgrowingsustainabilityconcerns.

AswemonitorthecarbonfootprintofourAImodels,werealisetheefficiencyofusingSLMstrainedonmorelimiteddatasetsforspecific,restrictedusecases,ratherthandeployingLLMsforeveryapplication.

CDOOFAMANUFACTURINGCOMPANY,NZ

Purpose-builtmodels,includingthosedesignedforlocallanguagesandnuancedregionalcontexts,willbeparticularlyindemand.Thesemodelsnotonlyaddressdiverselinguisticneedsbutalsoenhanceexplainabilityandarewell-suitedfordeploymentonsmalleroredgedevices.

ECOSYSTMAPACAIOUTLOOK2025|IBM

AsorganisationsrefinetheirAIstrategies,techanddataleaderswillevaluatemodelsbasedon:

GovernanceConstraints

Industrieswithstrictprivacyandsecurityrequirementsmayprefermodelsdeployableonisolatednetworksorcompliantwithspecificregulations.

TaskComplexity

Simplertasksmayonlyrequiresmallermodels,whereascomplex,

data-intensivetasksmightrequirelarger,moresophisticatedmodels.

DataAvailabilityandQuality

Smallermodelscanperformwellwithlimiteddata,whilelargermodelsoftendependonextensive,high-qualitydatasets.

ComputationalResources

TheavailabilityofGPUs,TPUs,andotherresourceswillguidemodelselectionandtrainingstrategies.

PerformanceMetrics

Latency,accuracy,cost-efficiency,andproximitytodatasourceswill

influencemodeldeployment,withmanyorganisationsoptingforedgecomputingtooptimiseinferences.

8

INTRODUCTION

THENEXTWAVEOFAI

SCALINGFORSUCCESS

CASESTUDIES

CASESTUDIES

AIModeltoSuitObjectives

KasikornBusiness-TechnologyGroup(KBTG)hasdevelopeditsownfoundationalLLMmodelcalled"THaLLE”(Text

HyperlocallyAugmentedLargeLanguageExtension),tailoredforfinanceandtheThailanguage.THaLLEhasachieved

CSALevel2certification,ensuringcompliancewithindustrystandardsforaccuracy,reliability,andfinancialanalysis.Byopen-sourcingthemodel,KBTGiscontributingtotheAI

communitywhileadvancingnext-generationAIgovernanceframeworks.

BangkokBank.ThefinancialindustryisrapidlyadoptingAItoenhancecustomerexperiencesandoptimiseoperations.BangkokBankchampionsthesynergyofhumansandAI

workingtogether,embracingacollaborativeintelligence

approachrootedinhuman-centredAIprinciples.ByleveragingAItoamplifyhumanpotential,thebankaimstoachieve

significantperformanceimprovements.Tounlockthefull

potentialofAI,thebankhighlightstheneedforindustry-

widecollaboration,responsibleinnovation,andasteadfastcommitmenttousingAIforthebettermentofsociety.

ECOSYSTMAPACAIOUTLOOK2025|IBM

9

INTRODUCTION

CASESTUDIES

THENEXTWAVEOFAI

SCALINGFORSUCCESS

#3UnifiedAI:EnsuringManagementandGovernance

In2025,evolvingregulations,diverseneeds,andresponsibleAIwilldriveorganisationstoinvestintoolsforvisibility,governance,andseamlessAIintegration.

AsAIevolves,dataandtechnologyleadersincreasinglyrelyonmulti-modal,multi-vendorenvironmentsthatintegratediversedatasourceslike

text,images,andaudiotopowerintelligentapplications.Organisationsmustnavigatecomplexregulations,suchastheEUAIAct,whilemanagingtheseintricateAIecosystems.Internaltechteamsfacethechallengeofensuringcompliance,fosteringresponsibility,andmaintainingtransparencyacrossmultipleAIsolutions.

Organisationswillconsider:

ModelOrchestration

HarnessingmultipleAImodelsrequiresrobustorchestrationtoolsto

manageandcoordinateworkflows,ensuringseamlessintegrationandpeakperformance.

VendorManagement

WithAIsolutionsfrommultiplevendors,organisationswilladopt

unifiedgovernanceframeworkstomaintainconsistency,security,andcompliance.Thiswillallowavendor-agnosticstrategytoenhance

flexibilityandadaptabilitytonewtechnologies.

DeveloperToolkits

Streamlinedtoolkitssimplifydevelopment,automatetasks,andenhancereliability.Featureslikeautomatedtesting,explainability,andintegration

withdiversetechnologiesenableorganisationstoaccelerateAIinnovation.

AutomatedAILifecycleManagement

Centralisedmodelinventorieswilltrackperformance,usage,andlineage,providingreal-timeoversight.Automatedmonitoringsystemswilldetectandaddressissueslikemodeldriftandperformancedegradation.

“AImodelsarepronetobiasanddrift,whichcanleadtounintendedconsequences.Tomitigatetheserisks,wecarefullycurateandmonitortrainingdataandregularlyevaluateandretrainmodelsforaccuracyandfairness.Automationisessentialtoachievethis.”

CTOOFABANK,SINGAPORE

ECOSYSTMAPACAIOUTLOOK2025|IBM

10

INTRODUCTION

CASESTUDIES

THENEXTWAVEOFAI

SCALINGFORSUCCESS

CASESTUDIES

PrioritisingSafeandSeamlessAIManagement

GSLab|GAVShasdevelopedZIF.AIthatexemplifies

governance-driveninnovationbyembeddingresponsible

AIpracticesintoitsGenAIsolutions.ZIF.AIhasenhanceditspredictiveandproactivecapabilitiestopreventapplicationandinfrastructuredowntimes.Keygovernancemeasures

includetransparencyinidentifyingLLMsources,robust

privacyfeatures,andastrongfocusonethicalAIpractices,

ensuringreliabilityandcompliance.ThesesafeguardsprovideguardrailsagainstissueslikeAIhallucinationwhilesupporting

proactiveissuedetectionandimproveddataintegrityassessments.

FeedloopAI,aleadingIndonesianGenAIprovider,has

partneredwithIBMtointegrateitsFL1AILargeLanguageModel(LLM)withthewatsonxplatform.AsoneofthefirstIndonesian-languageLLMs,FL1enableslocalgovernmentsandbusinessestomanageAIwithrobustgovernance,

risk,andcompliancetools.Throughwatsonx,Feedloop

customerscanautomateregulatoryobligationtracking,

ensuringcompliancewithcurrentandfuturestandardswhilemaintainingadherencetobusinessrequirements.

ECOSYSTMAPACAIOUTLOOK2025|IBM

11

INTRODUCTION

CASESTUDIES

THENEXTWAVEOFAI

SCALINGFORSUCCESS

#4AgenticAI:EmpoweringIntelligentSystems

AItoautonomouslyexecutetasksanddrivebusinessvalueasworkfloworchestrationbecomesincreasinglyessential.

Traditionalautomationtools,likeRPA,haveproveneffectiveinstreamliningrepetitivetasks.However,theyoftenstrugglewiththecomplexityanddynamismofreal-worldworkflows.Agenticworkflows,poweredbyAIagents,offeramoreadvancedandflexibleapproach.

"WeknowthatincorporatingAIrequiresredefiningourworkflows.However,traditionalworkflowsarecomplexandconsumevaluableresourcesaswenavigatesystems,copy-pastesequences,andhandleauthentication

hoops."

CDOOFATELECOMPROVIDER,INDIA

CombiningAIwithautomationdrivessignificantgainsinoperationalefficiency,customerexperience,anddecision-making.AsAIadvances,agenticworkflowswillbepivotalinredefiningthefutureofwork.

ECOSYSTMAPACAIOUTLOOK2025|IBM

WhyorganisationswillinvestinsolutionswithAgenticAIcapabilities:

Autonomy

AIagentscanindependentlyexecutetasks,makedecisions,andadapttochangingcircumstances.

Intelligence

LeveragingGenAI,theycanunderstandcomplexinstructions,reason,andlearnfromexperience.

Collaboration

AIagentscancollaboratewithhumanworkers,augmentingtheircapabilitiesandimprovingefficiency.

Adaptability

Agenticworkflowscanadapttochangingbusinessneedsandunexpecteddisruptions.

12

INTRODUCTION

THENEXTWAVEOFAI

SCALINGFORSUCCESS

CASESTUDIES

CASESTUDIES

AutomatingWorkflowOrchestrationforAIEfficiency

SirirajPiyamaharajkarunHospital(SiPH)hasrevolutioniseditspathologydiagnosticsthroughworkflowautomation.Byintegratinglaboratorysystems,imagescanning,andcentraldataprocessing,SiPHhassignificantlyimprovedefficiencyandaccuracyincancerdiagnosis.Thesystem’sautomatedworkflowsandAI-drivenslideimageanalysis—currently

pilotedforprostatecancer—streamlinestheidentification

ofpotentialcanceroustissues,allowingdoctorstofocuson

high-riskcases.ThistransformationlaysafoundationforfutureadvancementsincomputationalpathologyandAIdiagnosticsinThailandandbeyond.

AglobalupstreamoilandgascompanyisleveragingAItoautomateworkfloworchestrationandenhanceefficiency,

particularlyinseismicloganalysis.Byimplementingmachinelearningmodels,thecompanyautomatesdatacleaningandgap-fillingprocesses,significantlyreducingthemanualeffort

required.Thisenablesengineerstoprocess10-20logsin30

minutes–ataskthatpreviouslytookafulldayperlog-freeingthemtofocusonhigher-valuetasks.Theseadvancementsnotonlyimproveefficiencybutalsoenhancedecision-makingandproductivitybyprovidingfaster,moreaccuratedatainsights.

AsthecompanytransitionstooperationalisingAI,itcontinuestorefineworkflowstomaximisethetechnology'spotential

whilefosteringcross-functionalcollaboration.

ECOSYSTMAPACAIOUTLOOK2025|IBM

13

INTRODUCTION

CASESTUDIES

THENEXTWAVEOFAI

SCALINGFORSUCCESS

#5BeyondProductivity:TheHuman-CentricFutureofAI

In2025,organisationswillshifttheirfocusfrommerelyadoptingAItoolstoharnessingtheirpotentialforhuman-centredinnovation.WhileproductivitytoolshavebeenamajorfocusofAIadoption,thefutureliesinleveragingAItoenhance

humanexperiencesandcapabilities.

Foremployees,AIwillbecomeapowerfultooltoaugmenttheirroles,automateroutinetasks,andunlocknewopportunitiesforcreativityandinnovation.OrganisationswillprioritiseemployeeeducationandtrainingtoensureasmoothtransitiontoanAI-poweredworkplace.

Customershavebecomeaccustomedtochatbotinteractionsoverthepastfewyears.In2024,therewasapromisetoenhancecustomer

experiencesbyintegratingGenAIandadvancedcustomerintelligenceintotheseinteractions.Movingforward,human-centredAIdesignwillbeparamount.Byprioritisingempathyandengagement,organisationscanfosterstrongercustomerrelationshipsandbrandloyalty.AIsolutionswillbetailoredtomeetspecificcustomerneedsandpreferences,deliveringpersonalisedexperiencesthatdrivesatisfaction.

Byprioritisinguserneedsandpreferences,organisationscancreatemorepersonalisedandintuitiveinteractions,forbothemployeesandcustomers.Thisinvolves:

PersonalisedExperiences

TailoringAI-poweredexperiencesto

individualusers,

deliveringhighly

relevantcontentandrecommendations.

EmpatheticDesign

Understandinguser

emotions,motivations,andpainpointsto

designAIsolutionsthatresonate.

TransparentandExplainableAI

Providingclear

explanationsforAI-generatedoutputstobuildtrustand

confidence.

EthicalAI

EnsuringthatAIsystemsarefair,

unbiased,andalignedwithhumanvalues.

Continuous

Improvement

IterativelyrefiningAIsystemsbasedon

userfeedbackand

performancemetrics.

ECOSYSTMAPACAIOUTLOOK2025|IBM

INTRODUCTION

CASESTUDIES14

THENEXTWAVEOFAI

SCALINGFORSUCCESS

CASESTUDIES

Ahuman-centricapproachtoAIsuccess

AleadingASEANautomotivecompanyfacedsignificant

organisationalreluctancetoadoptAI,despitemanagement'sstrongbeliefinitspotentialasabusinessdifferentiator.To

tacklethis,thecompanydelayedimplementation,focusing

onfosteringacceptanceandaligningculturalshiftswithAI

integration.Keyinitiativesincludedupskillingandreskilling

employeestomeettheevolvingdemandsofthetechnology.

TheSouthWaikatoDistrictCouncil(SWDC)adopted

ahuman-centredAIapproachtoimproveinformation

accessibilityforitscitizens.Thecouncilimplementeda

virtualassistantenablingquickandaccurateresponsestouserqueries.With91.5%accuracyacrosstestquestions,thesolutionenhanceduserexperiencethroughnatural

languageprocessing,advanceddatafiltering,andtransparentaccesstodocumentsources.Thisinitiativenotonly

addressedinformationsilosbutalsoempoweredcitizens

withstreamlined,conversationalaccesstovitalinformation,exemplifyingAI'sroleinfosteringbetterpublicservices.

ECOSYSTMAPACAIOUTLOOK2025|IBM

15

INTRODUCTION

CASESTUDIES

THENEXTWAVEOFAI

SCALINGFORSUCCESS

AIacrossAsiaPacific:AComparison

ANZ

KeyDriversofAIAdoption

53%49%34%

Needtoreducecostsandautomatekeyprocesses

Pressurefromcustomers

Legalandregulatorycompliancepressures

KeyChallengesofAIAdoption

55%Costofimplementation/solution47%Ethicalconcerns

38%Limitedusecasesdefined

BiggestFocusofAIInvestmentsin2025

20%19%18%

Employeeexperienceandproductivity

Back-officebusinessprocessautomation

Salesautomationandcustomerlifecyclemanagement

N=202;Source:Ecosystm,2024

KeyDriversofAIAdoption

62%60%47%

Needtoreducecostsandautomatekeyprocesses

AdvancesinAIthatmakeitmoreaccessible

AIembeddedintostandardoff-the-shelfbusinessapplications

KeyChallengesofAIAdoption

46%42%38%

Dataaccessibilityissues

LimitedAIskills,expertise,orknowledge

Difficultyinintegrationandscaling

INDIA

BiggestFocusofAIInvestmentsin202527%Customerexperience

16%Planningandstrategy

16%OptimisationofITfunctions

N=216;Source:Ecosystm,2024

ECOSYSTMAPACAIOUTLOOK2025|IBM

16

INTRODUCTION

CASESTUDIES

THENEXTWAVEOFAI

SCALINGFORSUCCESS

SINGAPORE

KeyDriversofAIAdoption

48%31%31%

Competitivepressure

Environmentalpressures

Needtoreducecostsandautomatekeyprocesses

KeyChallengesofAIAdoption

45%39%33%

Limitedusecasesdefined

LimitedAIskills,expertise,orknowledge

LackoftheabilitytoproperlygovernAImodels

BiggestFocusofAIInvestmentsin2025

32%17%15%

Back-officebusinessprocessautomation

Planningandstrategy

Employeeexperienceandproductivity

N=84;Source:Ecosystm,2024

KeyDriversofAIAdoption

49%41%39%

Labourorskillsshortages

Needtoreducecostsandautomatekeyprocesses

Competitivepressure

KeyChallengesofAIAdoption

51%45%39%

Dataaccessibilityissues

LackofAIstrategy

LimitedAIskills,expertise,orknowledge

MALAYSIA

BiggestFocusofAIInvestmentsin2025

44%17%14%

Customerexperience

Salesautomationandcustomerlifecyclemanagement

Back-officebusinessprocessautomation

N=71;Source:Ecosystm,2024

ECOSYSTMAPACAIOUTLOOK2025|IBM

17

INTRODUCTION

CASESTUDIES

THENEXTWAVEOFAI

SCALINGFORSUCCESS

INDONESIA

KeyDriversofAIAdoption

56%37%33%

Needtoreducecostsandautomatekeyprocesses

AdvancesinAIthatmakeitmoreaccessible

Competitivepressure

KeyChallengesofAIAdoption

48%47%40%

LackofAIstrategy

Difficultyinintegrationandscaling

Limitedusecasesdefined

BiggestFocusofAIInvestmentsin2025

21%20%20%

OptimisationofITfunctions

Salesautomationandcustomerlifecyclemanagement

Planningandstrategy

N=81;Source:Ecosystm,2024

KeyDriversofAIAdoption

42%41%39%

AdvancesinAIthatmakeitmoreaccessible

Environmentalpressures

Pressurefromcustomers

KeyChallengesofAIAdoption

41%38%34%

Vendorlock-ins

Lackoftools/platformsfordevelopingAImodels

Costofimplementation/solution

THAILAND

BiggestFocusofAIInvestmentsin2025

29%18%16%

Back-officebusinessprocessautomation

OptimisationofITfunctions

Salesautomationandcustomerlifecyclemanagement

N=76;Source:Ecosystm,2024

ECOSYSTMAPACAIOUTLOOK2025|IBM

THENEXTWAVEOFAI

SCALINGFORSUCCESS

PHILIPPINES

KeyDriversofAIAdoption

47%47%45%

Needtoreducecostsandautomatekeyprocesses

Competitivepressure

Labourorskillsshortages

BiggestFocusofAIInvestmentsin2025

23%18%17%

Customerexperience

Back-officebusinessprocessautomation

Employeeexperienceandproductivity

KeyChallengesofAIAdoption

43%Limitedusecasesdefined

40%

Difficultyinintegrationandscaling

37%LackofAIstrategy

18

INTRODUCTION

CASESTUDIES

N=60;Source:Ecosystm,2024

ECOSYSTMAPACAIOUTLOOK2025|IBM

THENEXTWAVEOFAI

SCALINGFORSUCCESS

ScalingforSuccess:OvercomingBarriers

Despitepromisingearly-stageprojects,regulatoryhurdles,businessreadinessissues,andtechnologicallimitationscontinuetohinderwidespreadAIadoption.

NavigatingtheComplexitiesofAIImplementation

LeadingtheAICharge:CEOImperatives

DefineAI’sValueProposition.

Focusonhigh-impactAIusecases

alignedwithbusinessgoals,setclearfinancialobjectives,and

measureROIintime,cost,andoutcomes.Developascalable,sustainableroadmapforAI

initiatives.

AddresstheHumanFactor.

BridgetheskillsgapwithAI/

MLtalentdevelopment,manage

changetoovercomeresistance,

andempoweremployeeswith

clearcommunication,training,andsupport.

PromoteCross-Functional

Collaboration.AlignbusinessandITteamsunderaunifiedAIvision,governancemodel,andcentralisedAICentreofExcellence(CoE).

Secureexecutivesponsorshiptodriveprioritiesandresourceallocation.

StrengthenGovernanceandTransparency.Embedrobust

centralgovernanceand

transparencyintotheculturewhileempoweringbusinessunitswith

projectownershipforethicalandsecureAIpractices.

19

INTRODUCTION

CASESTUDIES

ECOSYSTMAPACAIOUTLOOK2025|IBM

THENEXTWAVEOFAI

SCALINGFORSUCCESS

BuildingaData-CentricOrganisation:GuidanceforDataLeaders

ConductaThorough

DataAudit.Address

qualityissuesand

leverageautomation

toolsforfastercleaningandannotation.Ensureaccesstosufficient,

high-qualitydata,

includingsyntheticdatawhenneeded.

PrioritiseDataPrivacyandSecurity.Preparetomeetamorecomplex

regulatoryenvironment,

beyondexisting

frameworkssuchas

theGDPR,DPDP,

PDPA,andthePrivacyActs.Establishinternalgovernancepolicies

andregularlyupdatesecurityprotocolstomitigaterisks.

AddressDataAccessandIntegration

Challenges.Break

downsilostoenable

seamlessdatasharinganduseintegration

toolstoharmonise

diversedatasources.

Strengthengovernancepracticesfor

consistencyandquality.

ClarifyDataOwnership

andIntellectualPropertyRights.

Defineclearpolicies

fordataownershipandintellectualproperty,

consultinglegalexpertstoensurecompliance

withregulations.

BuildaFuture-Proof

DataFoundation.WorkwiththeCIOtoupgradelegacysystems,adoptcloud-basedsolutions,andredefinethedata

architecturethatalignswiththeorganisation’sAIroadmap.

20

INTRODUCTION

CASESTUDIES

ECOSYSTMAPACAIOUTLOOK2025|IBM

21

INTRODUCTION

CASESTUDIES

THENEXTWAVEOFAI

SCALINGFORSUCCESS

TechLeadership