PWC

Innovatesmarter

HowAIistransforming

ResearchandDevelopment

1

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC2

Tableofcontents

ExecutiveSummary3

1Behindthecurtain:Today,srealityinR&Dandoperations4

Industrychallenges

5

TheAIrevolutioninR&D6

2AItosupportwithintheproductlifecycle7

Keyvaluableusecaseswithintheproductlifecycle,availabletoday8

BuildinganAI-enabledR&Dorganisation14

3SuccessfulAIimplementationsarebusiness-led,nottech-led17

4Conclusion20

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC3

ExecutiveSummary

Asindustriesfaceincreasingpressuretoinnovateswiftlyandsustainably,integratingartificial

intelligence(AI)intoResearchandDevelopment(R&D)processeshasbecomeessential.AIhasthepotentialtorevolutioniseR&Dbyaddressingchallengeslikeacceleratedtimetomarket,complexproductspecifications,andstrictregulatoryrequirements.

Bytransformingvastdataintoactionableinsights,AIenablesorganisationstostreamline

development,enhanceresourceefficiency,andbolstercompliance.Keyapplicationsthatare

achievablewithtoday’sstateoftechnologyincludevariantmanagement,requirementsengineering,andregulatoryalignmentthroughouttheproductlifecycle.EmbracingAIrequiresstrategic

adjustments,focusingondataquality,security,availability,andempoweringtheworkforcewithnewcapabilities.

ThePwCframeworklaysthegroundworkforovercomingimplementationbarriersandfosteringacultureofcontinuousinnovation,positioningorganisationsforlong-termsuccessinanincreasinglysustainability-focusedmarket.

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC4

1

BehindthecurtainToday’srealityin

R&Dandoperations

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC5

Industrychallenges

Asthe2025studyfromPwCincollaborationwithMicrosoft

AIin

operations:Revolutionisingthemanufacturingindustry

has

shown,thediscretemanufacturingindustryisunderimmense

pressuretoinnovate.Rapidlyevolvingcustomerexpectationsandtechnologicaladvancesmeanthatproductsandprocessesmustbeimprovedcontinuously.Innovationisnolongeroptional;itis

essentialfordifferentiation.

Timetomarkethasbecomeadecisivefactorincompetitiveness.

PwC'sresearchfromanupcomingstudyonthefutureofR&D

showsthatcompaniesmustacceleratetheirdevelopmentcyclestoseizeopportunitieswhilebalancingspeedwithqualityand

compliance.Meanwhile,growingproductcomplexityand

portfoliodiversificationarechallengingtraditionalR&Dprocesses,necessitatingmoresophisticatedcoordinationbetween

engineering,designandsupplychainfunctions.

Compoundingthesepressuresisthescarcityofskilledtalent.

Thereishighdemandforengineers,designers,andspecialistsinadvancedmanufacturing,whichlimitsthecapacitytoscale

innovation.Theissueofsustainabilityaddsanotherlayerof

complexity.Organisationsarecompelledtodesignproductsandprocessesthatreduceenvironmentalimpact,optimiseresourcesandcomplywithemergingregulatoryandsocietalexpectations.

Together,thesefactorscreateachallengingenvironmentinwhichincrementalimprovementsarenolongersufficient.Companies

mustadoptnewapproachesthatenhanceR&Deffectivenessandaccelerateinnovation.

Atthesametime,AIpresentsauniqueopportunity.Oursurveyreport,

AIinoperations:Revolutionisingthemanufacturing

industry

,producedtogetherwithMicrosoft,showsthatartificialintelligencefostersinnovationinbusinessenvironmentsby

optimisingdataanalysis,detectingpatternsandtrends,and

empoweringinformeddecision-makingandcreativesolutions.

Nearly60%ofrespondentsexpectsanincreaseinoperatingprofitmarginthroughtheuseofAI.

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC6

TheAIrevolutioninR&D

R&Disthecornerstoneofinnovationandattheheartofthese

challenges,makingitanidealareainwhichtoapplyartificial

intelligence.AIhasthepowertotransformthewayorganisationsinnovate,turningdataintoactionableinsightsandacceleratingdevelopmentanddecision-makingprocesses.

Byanalysinglargeandcomplexdatasets,AIenablesfasterconceptvalidationespeciallyforcomplexsystems,predictivemodelling

andoptimiseddesignprocesses,reducingrelianceoncostlyphysicalprototypes.AIalsohelpstoidentifyrisksearlyon,minimiseinefficienciesandalignproductdevelopmentmoreaccuratelywithmarketdemand.

Thesustainabilitybenefitsareequallycompelling.AIcansupportthecreationofresource-efficientdesignsandtheoptimisationofmaterials,aswellasthedevelopmentofcircularproduct

strategies,therebyembeddingenvironmentalresponsibilityintoR&Dfromtheoutset.

IntegratingAIintoR&Disastrategicshift,notmerelya

technologyupgrade.Itstrengthensinnovationcapabilities,shortensdevelopmentcyclesandenablesmanufacturerstocompeteeffectivelyinafast-moving,complexand

sustainability-drivenmarket.

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC7

2

AItosupportwithintheproductlifecycle

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC8

Keyvaluableusecaseswithin

theproductlifecycle,availabletoday

Intoday'sdynamicbusinessworld,theuseofAIisbecoming

increasinglyimportanttoenhancetheefficiencyand

competitivenessofcompanies.Particularlyinthecontextoftheproductlifecycle,AIopensamultitudeofopportunities.

Productlifecycle

Innovation

Productdevelopment(e.g.V-model)

Order

Realisation

process

Phaseout

Idea

management

Projectscoping

Project

feasibility

Productstrategy

ProductDraftLaunchSellProduce,

conceptionfreezedeliver

andservice

Endoflife

Process

Eachofthefivemainphasesoftheproductlifecycle—innovation,productdevelopment,realisation,orderprocess,andphase-out—presentschallengesthataresolvablewithtoday'sstateofAI

technology.Thesetechnologiesempowercompaniestostreamlineoperations,anticipatemarketneeds,andswiftlyadapttochange.Itpresentsanopportunitytoacceleratetimetomarkettimelines,improveproductsbycostandinnovation,easeworkloadfor

resources,improveR&Dpowerbyfacilitatingcollaboration,alignmentandcoordination,tonameonlyafewofAI’svaluedrivers.ThefollowingsectionspresentmatureusecasesforAIwithintheproductlifecycle.

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC9

AItooptimisevariantand

complexitymanagement

Inproductdevelopment,masteringvariantmanagementisoftenchallengingduetoitsinherentcomplexity,necessitating

innovativesolutionsinR&D.Thekeyliesineffectivelymanagingexternalandinternalcomplexities—balancingportfolio,module,andcomponentvariantswithcustomerrequirements.Ratherthaneliminatingcomplexity,thegoalistoharmonisemarketneeds

withcompanyofferings.Utilisingrealdata,likesalesfigures,iscrucialtoquantifyingandvaluingthiscomplexity.

Effectivevariantmanagementdistinguishes‘highrunners’from

‘lowrunners’andlinksthesetotechnicalimplications,suchasthenumberandseverityofcomponentvariantsrequired.Misjudgingthecostofcomplexitycanleadtofinancialinefficienciesandover-engineering.Manycomplexitymanagementsolutionsareoften

impracticalandfurthercomplicatematters.

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC10

CurrentAIsolutionstailoredforR&D,likePwC’sMETUS*,address

thesechallengesbyoptimisingproductconfigurationsusing

advancedalgorithms.Theyhelpreduceunnecessarycomplexity,minimisecomponentcounts,andmaintainawiderangeof

configurationoptions.METUSleveragesAItoquantifytheimpactofexternalvarietyonaportfolio,evaluatetechnologyoptions,andproposebalancedsolutions.Drawingoninsightsfromourclient

engagementsandexperience,thisleadstoresourceefficiencies,achievingupto50%variantsavings,25%componentreductions,andupto33%costsavings—enhancingagilityandmarket

responsiveness.

Supply

chain

Market

Fit

L_」

LLM

withliveconnectiontofullyconnectedMETUSdatamodel

A

METUS

APwCProduct

Servicestructure

productstructure

AItodriverequirementsengineering

Requirementsengineeringprovidesthefoundationforproductdevelopment.Weakrequirementsengineeringcanslowprogressandincreaserisk.Itcanleadtofragmentedcommunication,

incompleteorinconsistentdocumentation,andscopemisalignment—sometimesevencommercialfailures,whenproductsmissthemarkwithcustomersorthemarket.

*PwC’s

METUS

isanadvancedmethodologyandsoftwaresuitedesignedtooptimiseproductdevelopmentandportfoliomanagement.LeveragingAI,METUSenablesorganisationstosystematicallymanagetheirproductsandservices.Theplatformsupportsend-to-endandcross-departmentaldigitalmodelingandintegratesseamlesslywithPLM/ERPsystems.Formoreinformation,visit:

https://pwc.to/3JEzMTH

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC11

Operationally,poorrequirementsmanagementdrivesdelays,cost

overruns,andresourceinefficiencies.Misalignmentamongteamsandalackofclarityonprojectobjectivesfurtherexacerbatethesechallenges,compromisingtheproduct'squalityandalignment

withclientneeds.

GenerativeAI

Machine-readabletext

OpticalCharacterRecognition(OCR)

+

Atomic,clear

andconsolidatedrequirements

Semanticrules

Handwrittenrequirementsspecification

Commonunderstandablerequirementsforthe

downstreamprocesses

GenerativeAIaddressestheseissuesbyenablinginstantextractionandconsolidationofrequirementsusingsemanticsandoptical

characterrecognition(OCR)totranslatehandwritteninformationintomachine-readabletext.Ourworkwithclientsdemonstrates

thatAI-drivensystemscanreducerequirementsderivationtimeby~60%,documentationeffortbyupto30–40%anddecrease

reworkby~25%throughimprovedrequirementconsistency,resultinginfasterdevelopmentcyclesandbetteralignmentbetweendesignoutcomesandcustomerexpectations.

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC12

AItofacilitatecomplianceaspartofR&D

Intoday’sworld,thefunctionsofR&Dandcompliancearebecomingincreasinglyconnected—evenintegrated.Thisisduetocontinuouslyincreasingregulatoryrequirementsthatcausecompliancetobecomeanintegralpartofphaseswhereproductsarecreatedandmodified.Currentchallengesinproductcomplianceincludehighmanualefforttointerpretcomplexregulations,evaluatingand

mitigationofriskofnon-compliancefrominconsistentimplementation,andtime-intensiveauditpreparation.AItechnologycansupportinthesetasksalready.

OurexperienceshowsthatanAI-poweredcomplianceassistantisaneffectivesolution.Itidentifiesglobalregulatorychanges,translatesthemtoyourportfolio,andhelpsevaluatenecessary

modificationsacrosssystemsandsub-systems.Wehaveseensuchtechnologyeffectivelyspanfromautomaticscanningtoevaluationofcurrentcompliance,throughtoconnectingtoR&Ddepartmentstoexecuteneededchanges.Ittherebystreamlinestheprocessandreduceshumanerror.ThebenefitsofthisAI-integratedsystemaresubstantial,itachievesapproximately50%timereduction,upto60-70%reductioninmanualeffort,and~40%errorreduction.Theseimprovementsnotonlyreduce

auditcostsbutalsoenhanceaccuracyandefficiency,fosteringamorereliableandcompliantR&Denvironment.

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC13

AItoboostdatamigration

Datamigrationprogrammesoftentakeseveralyearstoexecute.InR&D,productlifecycle

management(PLM)transformationsarecommon—andtheytypicallyinvolvemigratinglarge

volumesofdata.Frequently,weseeasignificantvolumeofresourcesbeingdedicatedforthesole

purposeofsuchdatamigrations.Traditionally,itisoftenaverymanualprocessfordatatobe

validated,exportedandimportedfromdifferentdatadomainswithintheproductdevelopment

process(e.g.,CAD,BOMs).Companiesoftendealwithalargevolumeofdatatobehandledandseekspecialistexpertiseandconsistentinvolvement,makingtheentireendeavourtime-consumingand

expensive.

AI-basedsolutionswithtoday’sstateoftechnologycanautomatedatamigrationandvalidation

processesalready.Theycansupportinextractingdata,integratingdocuments,andcross-checkingvalidationresults.Keybenefitsseeninourprojectsworkincludetimesavingsofupto60%throughacceleratedtaskexecution,potentialcostreductionsofapproximately50%inprojectdelivery,andthoroughdataprotectionthatrequiresonlyabout25%involvementfromspecialiststaff.The

implementationofAI-drivenefficienciesenhancesbothlarge-scaledatamigrationandadaptivevalidation.

Datarequiredforvalidationprocess

?windchilli

TargetPLMsystem

1

Datarequired

forvalidationprocess

Extracteddatadumps

BMIDEDataModel*

5DataMigrationTool

Correctsourcedata

accordingtovalidationresults

Rawdata

Palantir

Datasetsandrelationships

Datatobevalidated

Resultsandfeedback

Importandpreparesourcedataforvalidation

Ontology:organiserelationshipswithinthedata

AI-powereddatavalidation+humanmonitoring

4

2

3

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC14

BuildinganAI-enabledR&Dorganisation

TheimplementationofAIpresentsimmensepotentialand

transformativepoweracrossvariousindustriesandbusiness

processes,asanalysedindetailbyPwCandMicrosoft’s

AIin

Operations

study.However,despitethetechnological

advancementsandthepromisedbenefitsofAI,organisationsfaceamultitudeofchallengesthathindersuccessfuldeploymentandutilisation.Thefollowingchartshowsanumberofnoteworthy

factorsthatarewidelyprevalentanddeeplyentrenchedin

operationalandstrategicprocesses.Wediscusseachoftheseinmoredetailinthenextsection.

BiggestchallengestoimplementAI42.4%

23.7%23.2%22.3%

19.9%18.5%

DataqualityITanddatasecurityDataavailabilityCostofAIsoftwareTechnologymaturityLackofAIknowledge

concernsandinnovationspeedacrosstheworkforce

Source:

PwCandMicrosoft,AIinoperations:Revolutionisingthemanufacturingindustry

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC15

Whyclean,consistent

dataiscriticalforAIsuccess

Safeguardingsystemsinanincreasingly

connectedworld

Breakingdownsilostounlockusabledata

Balancinginnovationwithinvestment

DataformsthebedrockofanyAIimplementation,anditsqualityfundamentallydictatestheefficiencyandperformanceof

algorithms.InR&D,dataqualityissuesoftenarisefrom

inconsistenciesintest-benchmeasurementsandundocumented

designiterations.LegacyPLMandsimulationsystemsmaystore

resultsinincompatibleformatsorlackstandardisedmetadata,

complicatingcross-projectanalyses.Consequently,AImodels

trainedonsuchfragmentedanderror-pronedatasetscanyield

unreliablepredictionsandlimitthediscoveryofmeaningfuldesigninsights.

ProtectingsensitiveinformationandestablishingdataintegrityareparamountwhenimplementingAIsystems.Theincreasing

integrationandinterconnectivityofsystemselevatetheriskofcyber-attacksanddatabreaches.Organisationsmustimplementrobustsecuritymeasurestocombatthesethreats.Thischallengearisesfromacontinuouslyevolvingthreatlandscapeanda

frequentlyinsufficientpreparednesstomeetnewsecurityrequirements.

Anothersignificantchallengepertainstotheavailabilityofthe

requiredvolumeanddiversityofdatanecessaryfortrainingAI

models.Often,theessentialdataeitherdoesnotexistinthe

desiredformatorisdifficulttoaccess.Thischallengeisdeeply

rootedinhistoricallydevelopedinformationsilosandproprietarydatabasesthathindersmoothdataflow.Companiesmust

strategisetocollectandprovidedatamoreefficiently.

InvestinginAItechnologiescaninvolvesubstantialfinancialandhumanresources.Thecostsassociatedwithdeveloping,deploying,andmaintainingAIsystemsposeasubstantialeconomichurdleformanyenterprises.Thischallengeisfrequentlylinkedtothe

necessityofacquiringspecialisedsoftwaresolutionsandengagingprofessionalsforsystemoversightanddevelopment.

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC16

Keepingpacewith

rapidAIevolution

Buildingcapabilitytodriveadoption

TherapidlyevolvinglandscapeofAItechnologybringsforthbothopportunitiesandchallenges.Organisationsfinditdifficulttokeeppacewiththeinnovationspeedandcontinuallyupdatetheir

systems.Thischallengeoriginatesfromthenatureofthe

technologyitself,asnewbreakthroughsandimprovementsoccurinrapidsuccession,renderingexistingsystemsandprocesses

quicklyobsolete.

AnotherprominentbarrieristhegenerallackofknowledgeandskillsrelatedtoAIwithintheworkforce.TointegrateAI

successfully,knowledgeofthetechnologymustbewidely

disseminated,necessitatingtargetedtrainingandeducation.TheoriginofthischallengeliesintherelativelynovelnatureofAI

technologiesandtheshortageofestablishededucationalframeworks.

ThesechallengeshighlightthecomplexityofAIimplementationandelucidatehowdeeplytheir

causesandprevalencearerootedwithincorporatecultureandexistingtechnologies.Astrategic

approachencompassingbothtechnologicalandorganisationalsolutionsisimperativetoovercomethesebarrierssuccessfully.

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC17

3

SuccessfulAI

implementationsarebusiness-led,nottech-led

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC18

Abusiness-ledapproachisnecessarytoalignAIinitiativeswith

organisationalstrategyeffectively,integratingAIintoa

comprehensiveframeworkthatencompasseskeydimensions—strategy,products,processes,IT-technology,andorganisation.

ThissynergisticapproachbeginswithinspiringanddefiningthevisionforAIadoption,ensuringalignmentwiththeorganisation'sstrategicobjectivesandhigh-levelbusinessambitions.

Initially,byfocusingonaligningAIcapabilitieswithstrategic

goals,organisationscanidentifyimpactfulusecasesandassess

readinessfrombothbusinessandtechnicalperspectives.This

alignmentiscrucialforensuringthatAIsolutionsenhanceexistingprocesses,drivinginnovationandcompetitiveadvantage.

PwCsupportsclientsfrominitialscopingtoend-to-endimplementationandbringsinacceleratorsineverystage

InspireandscopeAssessStrategiseDevelopScaleEnable

Inspirationand

scopingworkshop

Readiness

assessment

(Gen)Al

strategy

Pilot

usecases

Scalable

platform

Global

enablement

?Shareanoutside-in

perspectiveonGenAl.

?ShowcaseexemplaryusecasesonGenAl.

?Understandhigh-levelbusinessneedsandambition

?AssesstheGenAl

capabilitiesfroma

businessand

technicalperspective.

?Identifyfocusareasanddriveroadmap

?DevelopGenAl

aspirationandvision

?CollectGenAlusecasesandassessfeasibilityand

prioritizethem.

?Collectrequired

capabilitiestosupportGenAladoption

?Detailoutbusinessproblemand

understandbusinessprocess

?Derivetechnicalanddata-related

requirementsandarchitecture·

?ImplementMVP

solutionforselectedusecase

?DesignAlPlatform

architecturebuildingupontheData

Platform.

?BuildDataPlatform

?Setupuniform

DevOpsprocessesandgovernance

processes

?Designplatformoperatingmodel

?Developupskillinginitiativeanddefinetargetgroups

?Designtraining

conceptandexecutewithinorganization

?Setupchangemanagementcampaign

Provenworkshopformat

Readinessassessment

GenAlstrategy

framework

Alusecasecompass

Dataplatformlibrary

Dataand

Alacademy

~1week

~4weeks

~8weeks

~12weeks

~20weeks

~30weeks

Innovatesmarter:HowAIistransformingResearchandDevelopmentPwC19

Strategisingbecomesafocalpoint,wheredetailedplansare

craftedtosupportAIadoption.Thisinvolvesprioritisinghigh-

impactusecases,fosteringend-to-endcontinuityinprocesses,andcreatingscalablesolutionsthatintegrateseamlesslyintoexistingITframeworks.Theemphasishereisonenhancingoperational

efficiencyandenablingfastertime-to-marketthroughefficientengineeringpracticesandrobustdatainfrastructures.

Developmentofpilotusecasesfollows,wheretechnical

architecturesarespecifiedandMinimumViableProduct(MVP)

solutionsareimplementedtoyieldimmediatevalue.Thisphase

underscoresthepracticalapplicationofAI,demonstratingtangibleoutcomesthatinformbroaderintegrationstrategies,supportedbyseamlesstoolchainsthatenhancecapabilities.Toeffectively

developandintegrateAIusecases,partneringwithatechnologyproviderlikeMicrosoftofferssubstantialadvantagesanddeliversimprovedoutcomes.

Leveragingpreconfiguredservicesandestablishedbestpracticesreducesdevelopmentandimplementationcomplexity,thereby

acceleratingtimetovalue.Furthermore,thecollaborationenablestheapplicationofstate-of-the-artITsecuritycontrolstomitigatesecurityandcompliancerisksandensurethesecure,compliant

deploymentofAIacrosstheorganisation.

Finally,organisationalenablementisacriticalcomponentoftheAIapproach,reinf