亲爱的首席信息官，您是否充分利用了AI？您的AI战略设计选择

www.bernsteinresearch.com DETAILSINVESTMENT IMPLICATIONSWe like Indian asset management businesses for their earnings growth and high profitability. We think Al will enable betterAMC as Outperform and Nippon AMC as Market-Perform.Note that this is a non-exhaustive survey of Al tools and does not constitute a product endorsement or recommendation topurchase any of the tools identified.EVENT ALERTWe are hosting Pascal.Al, a startup that is building Al agents for institutional investors - 14 Jul 2025, 3:30pm IST (6:00pmHKT/11:OOamGMT/6:OOamET).Joinusforaninsightfulconversationwiththeco-founderofPascalAl-acustomizedlargelanguage model built specifically for institutional investors. Pascal is already working with over 20 funds worldwide, and has met300+ institutional investors on the topic.Register in order to receive the dial-in details: Click here to registerBUY-SIDE NEEDS AN AI STRATEGYThereare multipleAl-poweredtoolstryingtopitch and win mandatesfrom institutional investors.Giventhe nascent stages ofAl itself, many Al tools are springing up to service the asset management market. Incumbents already under contract with mostHaving seen some of the Al offerings available in the market evolve over the last few months, we see their potential to revamphow invetsors work on open-ended research questions (often dished out to sell-side today), idea generation, sector/s tockanalysis & portfolio tracking. We see Al tools as a wayof boosting productivity of investment teams, at the bare minimum.Enhanced productivity allows funds to evaluete mlore ideas, or do deeper diligence into select ideas. This should help drivebetter alpha-capture for the portfolio stecks.The bigger opportunity to our md comes from integrating publicly available data with your internal investment team datainto an Al-ready platforgtafoknown as 'Retrieval-Augmented Generation, or RAG-tools). Tools can today (unassisted, onceotherwise),and build/infercross-company insights.Think asector-levelround-up after earnings,but automaticYou can talk to these models, as if talking to your sector analyst - parse trends, see how market conditions are changing, anddraw a pecking order for your coverage names. All with minimal human effort/ involvement. Imaging all your expert call notes,research insights, management meeting notes are indexed and ready to be absorbed into replying to the next question you askthe Al engine. These notes may be created by you, your analysts, or your counter-part covering the same sector for a differentgeography for your fund.Al-ready indexing of investment team insights also creates an institutional memory, and reduce attrition-related risks. A seniorPM with decades of experience retiring, who specialized in a certain sector, does not create a void in the fund's knowledge base.It reduces talent dependencies, and creates a knowledge base that an incoming resource can reliably leverage to scale-up, andmaintain continuity in investment views.The good thing is that a growing set of CiOs are showing a realization of the potential of these tools, and are keen to explore.Our conversations withfoundersbuilding inthespace indicate CiOinteractions rangefrom'outrightdismissive'to‘keentoPMs/analysts-butmoreseniorfolks realizethe inevitable,and wanttopositionasthe internal championfor AlimplementationThis earnsthemanedgeonpeers intheirpositioning inthe internal/externaltalentpool.INDIA ASSET MANAGEMENT INSURANCE C BERNSTEIN SOCIETE CENERALE CROUP A TOP-DOWN VIEW ON DESIGN CHOICES FOR YOUR AI PARTNER SELECTIONOutside in, startups are seeing the asset management industry as a large B2B software market with an ability to pay for Al-powered tools that help expand investment team capacities. There are a lot of offerings, and investors are getting pitches frommultiple players. While investors are happy to take a pitch, and a trial, satisfaction levels vary linked to the product itself, andadoption curves within the fund. We try to de-clutter the decision, and share a top-down view of how investors should decidewhich Al tools/vendors to work with. We also share the key design choices involved in building your suite of preferred Al tools.Exhibit 1 summarizes our preferencesEXHiBiT1:Atop-downviewondesignchoicesforyourAlpartnerselectionWhat it means?Design PreferenceBuy/Build/PartnerPartner customizatiorPartnerCustomizationVertical v. HorizontalVertical LLMsLLMsvs. Horizontal LLMs jack of all tradesSingle-model v. Multi-modelbuilt on a particular LLMMulti-modelAl-native vs. Al-overlayIntegrate Al into existing solutions vs. buildAl-native solutionsAl-native solutionsPlatform longevityTry to partner with platforms that have abilityScaleedgeBernstein analysiBUY VS. BUILD VS. PARTNERLike any decision-making process, there is a buy v. build discussion when it comes to picking an Al tool/ platform.We think investors are best advised to subscribe to off-the-rack solutions as a starting point (low cost, good ROl, gauge usage/adoptionpatterns),buttolookatpartn