Agefi Luxembourg - octobre 2025

T obetter understandwhat theC- suite thinks about their organiza- tion’s use ofAI, its risks and impact on theworkplace, EYandFT Longitude conducted a “EuropeanFi- nancial ServicesAI Pulse Survey”. CompletedbetweenMarch and June 2025, the survey gathered responses from410 leaders in banking, insurance and wealth and assetmanagement, representing organizations with assets from$1 billion to over $1 trillion across 16 coun- tries.Among the respondents were leaders fromLuxembourg. Here are some of themost interes- ting takeaways fromthe research across Europe. Leaders are heavily investing inAI, but still not convinced they are prepared for tech-related risks Financial services firms are investing heavily in AI training (88% report moderate to extensive invest- ment),modeltestingandauditing(84%),anddataac- cess control (83%). Yet, over half (57%) of firms (and 60%ofwealthandassetmanagers)areconcernedthat their organization’s approach to technology-related risk is insufficient for emergingAI technologies. Notably,30%oforganizationshavenoorlimitedcon- trols to ensure AI is free from bias. And, while most firmshave some riskmitigationplans inplace, onlya littleoverhalfrelyoninternalaudits(52%)toprovide trust and confidence in theirAI systems, though ap- proachesvarybysectorandregion.Afterinternalau- dits,firmsemployconsultationwithindustryexperts andthird-partyAImodeltestingandvalidationasthe nextmostcommonmethodstoprovidetrustandcon- fidenceinAIsystems.Thechallengeliesinthescarcity ofAI literate resources and the effort needed to train theworkforce in a verynewanddynamic field. OrganizationsatmoreadvancedstagesofAImaturity (“Transforming” or “Leading”) feel better equipped tomanageAIrisks,butevenamongthese,halfbelieve theirapproachisstillinsufficient.Controlsarereport- edly strongest among banking and capital markets companies. Only a third ofwealthmanagers are comfortablewith agenticAI, but are using it anyway Over 40% of financial services firms (and 33% of wealthmanagers)areextremelyormoderatelyfamil- iarwithagenticAI(thecurrentstateoftheartforlarge languagemodels).Despitethis,already35%offinan- cialservicesfirms(andover40%ofwealthmanagers) say they are currently using it, while 25%plan to im- plement itwithin the next sixmonths. Consideringthebroaderrangeoffeatures,capabilities andpotentialusecasesofAI(e.g.,multimodalAI,syn- thetic data generation, quantum machine learning, autonomous robots, etc.), fewer than 50% of wealth managers(andfinancialservicesfirmsingeneral)are moderatelyorexceptionallyfamiliar.Interestingly,au- tonomousrobotsareexpectedtoseebroaderadoption over the next year. The latter being the intermediary steps before jumping into Agentic AI which will or- chestrate complexworkflows inorder todeliver per- sonal, efficient and scalable outcomes. Fears of job losses and less intelligent work Many leaders worry about AI’s potential to cause significant job losses, manipulate consumer percep- tions, and generate false information (e.g., deep- fakes). Concerns also extend to the negative impact on vulnerable groups in society. Wealth and asset managers specifically are more concerned thanbanks or insurers aboutAI resulting in significant job losses, that AI will be used to ma- nipulate how consumers think and feel and that AI will become uncontrollablewithout human insight. They also have the least trust in their consumers – only 32%of wealthmanagers agree that consumers trust that companies in their sector will manage AI in away that aligns best with their interests. Many C-suite executives fear that excessive depend- ence onAI coulddiminishworkforce cognitive abili- ties. There is also concern about accountability, transparency, ethics, data protection, cybersecurity and the potential for disinformation.The industry’s strict regulatory requirements and the highdata sen- sitivityaddalsomorepressureonWM’swhoarecon- cerned by reputational risks tied to opaque or biased AI-drivenfinancial decisions. HowcanAI be embraced whilemanaging for risk? Linkwith strategy For wealth and asset managers, embracingAI effec- tively begins with a clear strategy that linksAI initia- tives directly to business objectives. Rather than experimenting with AI in isolation, firms shouldidentifywhereautomation,predic- tive analytics, andgenerativeAI can create measurablevalue,fromimprovinginvest- ment research to enhancing client person- alization. Building a data foundation is critical here: high-quality, well-governed data ensures that AI models are accu- rate, auditable, and aligned with regulatory expectations. Lead- ership buy-in is equally im- portant, as executives must set the tone for howAI is inte- grated into decision-making and client offerings. Evolve frameworks with AI adoption Risk management must evolve in parallelwithAI adoption. Traditional risk frame- worksmay not fully capture the unique challenges of AI, such as model bias, explainability gaps and unintended consequences. Firms should establish cross-functionalAI governance committees that in- clude compliance, IT, investment professionals, and risk managers to evaluate potential impacts before deployment. Scenario testing, stress simulations, and ongoing monitoring can help identify vulner- abilities early, reducing the chance of reputational or regulatory fallout. Transparencywith clients and stakeholders about howAI is used is also a growing expectation and can build trust. Bring your people alongwith you Finally,firmsneedtobalanceefficiencygainswiththe human expertise that underpins the industry. AI should augment rather than replace skilled profes- sionals,allowingthemtofocusonhigher-valueactiv- ities like portfolio strategy and client relationships. Thenextfewyearswilllikelydefinewhichfirmssuc- cessfully harness AI to deliver better outcomes for clients.Thesefindingshighlightnotonlythegrowing investmentinAIacrossthewealthandassetmanage- mentindustry,butalsothepersistenttensionbetween innovation and risk, making it clear that firms must strike a careful balance as theymove forward. AbdelhayTOUDMA, EYLuxembourgPartner, Consulting,CyberandDigitalRisk JensSCHMIDT, EYLuxembourgPartner, Wealth&AssetManagementConsultingLeader Is it possible for wealth managers to embraceAI while managing the risks? BanquedeLuxembourg,sociétéanonyme–14,boulevardRoyal–L-2449Luxembourg–R.C.S.B5310 For over 40 years Banque de Luxembourg Asset Servicing has been helping initiators to set up their investment funds in Luxembourg. Our specialist teams are ready to assist you with tailor-made services, expertise and unwavering support. Supporting fund initiators since 1983 Octobre 2025 15 AGEFI Luxembourg Wealth management