The quiet momentum that changes everything

I used Claude (Anthropic) to help research and write this piece. The analysis and perspective are mine, but Claude did the heavy lifting on synthesizing industry data and making my draft readable.

The AI Inflection Point in Asset Management: An 18-Month Window

We’ve hit an inflection point. Oliver Wyman’s 2025 industry forecast identifies 2026 as a year of inflection for both the economy and asset management. Microsoft’s Bill Borden put it more bluntly: success in 2026 won’t come from experimenting with AI, but from re-architecting core business processes to be human-led and AI-operated.

After two years of pilots and proofs-of-concept, the question isn’t whether AI will transform investment management. It’s whether your firm will be among those capturing value or explaining why transformation stalled. The data shows we’ve crossed a threshold. What happens next separates firms building lasting competitive advantage from those still debating governance frameworks in committee.

The Paradox: Universal Adoption, Elusive Returns

The survey numbers look impressive until you dig into them. Grant Thornton’s 2025 survey of 500 asset management executives found 73% consider AI critical to their organization’s future, with 77% claiming an effective AI strategy. EY surveyed 100 wealth and asset managers and found 95% have scaled generative AI to multiple use cases.

But the returns tell a different story. Carne Group’s research shows fewer than one in five firms have deployed AI for core operations. Two-thirds report only small or moderate returns on their AI investments, with 12% seeing no returns or negative results. Publicis Sapient analyzed firms managing $74.2 trillion in assets and found only 19% report ROI greater than 7%. “Asset managers are talking a big game on AI, but the sector is still waiting for true transformation,” says Siobhan Noble, Chief Data and AI Officer at Carne Group. “The imperative is to first understand the problem statement, fix the process, and then apply technology.”

This gap between experimentation and production-scale deployment is the industry’s greatest challenge and its most significant opportunity. McKinsey estimates AI could eventually impact 25 to 40 percent of an average asset manager’s cost base, but only for those who move beyond pilots to enterprise-scale implementation. The problem isn’t technology availability. It’s that approximately half of all asset management firms lack basic processes to clean, normalize, and tag their internal data. Grant Thornton’s research ranked “data quality” as the second-most-cited barrier to AI success (51%), just behind cultural resistance (55%). Trying to bolt advanced AI onto legacy data architectures produces the modest returns that populate these surveys.

What Production-Scale Actually Looks Like

The firms achieving measurable business impact share a pattern: they’ve moved from treating AI as point solutions to embedding it as operational infrastructure. The contrast between industry averages and what leading firms have achieved is striking.

Morgan Stanley’s AI Assistant launched in September 2023 and now achieves 98% adoption across financial advisor teams. The tool accesses over 100,000 research documents and improved document retrieval efficiency from 20% to 80%. Critically, advisors report spending more time on client relationships because routine research tasks now take minutes instead of hours.
JPMorgan deployed its LLM Suite to approximately 50,000 employees (15% of its workforce). During April 2025 market volatility, the firm’s Coach AI enabled rapid personalized client responses, contributing to a 20% year-over-year increase in asset and wealth management sales. The firm also launched Proxy IQ, becoming the first major asset manager to replace third-party proxy voting advisors with AI-powered analysis.
UBS processes 8 million AI tool prompts per quarter across 52,000 employees, with over 300 active AI use cases. The firm digitized 60,000+ documents into a queryable knowledge base and decommissioned 1,100 legacy applications as AI-powered alternatives replaced fragmented workflows.
In the hedge fund space, the transformation is equally clear. Generative AI adoption among hedge funds grew from 6% before 2022 to 63% by 2024. Funds adopting these tools report 2-4% higher annualized abnormal returns compared to non-adopters. Point72’s AI-focused Turion Fund returned 14.2% in 2024 versus 6.2% for the Nasdaq Composite.

The applications gaining real traction fall into three categories:

Investment research automation represents the most mature area. Platforms like AlphaSense can query 450 million+ documents, while competitors like Hebbia and Rogo have moved from curiosity to essential infrastructure. Deloitte estimates top global investment banks could see front-office productivity improvements of 27 to 35 percent by 2026 from AI research tools. Tasks that previously consumed days (synthesizing earnings calls, updating financial models, analyzing 10-K changes) now complete in minutes.
RFP and DDQ automation has emerged as an unexpected high-impact area. Asset managers responding to institutional consultant queries face punishing timelines with enormous competitive stakes. New purpose-built solutions cut response times from weeks to hours. For firms completing hundreds of RFPs annually, the operational leverage is substantial.
Client engagement and distribution tools are accelerating rapidly. EY data shows 62 to 74 percent of asset managers now prioritize investments in personalized client outreach. One firm reported a 74% increase in win rates and 95% boost in quota attainment using AI-powered content management. The underlying issue: relationship managers spend 60 to 70 percent of their time on non-revenue-generating activities due to outdated systems. AI is reclaiming that time.

Agentic AI (systems capable of autonomous reasoning, planning, and action) represents the next frontier. While only 7 to 10 percent of firms have deployed agentic capabilities today, 78% are actively exploring them. Gartner predicts 33% of enterprise software will include agentic AI by 2028, up from less than 1% in 2024. Franklin Templeton’s strategic partnership with Wand AI signals how leading firms are positioning for this transition.

The 18-Month Window and What It Requires

This progression from research automation to agentic systems creates the urgency behind McKinsey’s timeline. McKinsey frames the competitive stakes as a “600-day imperative”: firms have roughly 18 to 24 months to move from proof-of-concept pilots to enterprise-scale AI deployment. Those who achieve this transition will capture cost and speed advantages that late adopters cannot match. Those who don’t risk finding themselves structurally disadvantaged as AI-native competitors and well-resourced incumbents pull ahead. This timeline is particularly unforgiving in asset management. Unlike industries where AI adoption can happen gradually across independent business units, asset managers face three compounding pressures. First, fiduciary duties and regulatory oversight mean firms cannot afford the experimental failures that other sectors tolerate - the governance and compliance infrastructure must be production-ready from day one. Second, institutional client relationships operate on winner-take-most dynamics: once competitors can deliver RFP responses in hours instead of weeks, or provide research synthesis in minutes instead of days, the expectations reset permanently. Third, the data infrastructure debt unique to asset management means firms cannot simply bolt AI onto existing architectures - they need foundational rebuilds that take time even when prioritized.

The firms that made early infrastructure investments are now entering the acceleration phase. The capital flowing into AI infrastructure reflects its strategic importance: Brookfield announced a $100 billion AI infrastructure program with NVIDIA and Kuwait Investment Authority.

The Stargate Initiative represents $500 billion in commitments from OpenAI, SoftBank, Oracle, and Microsoft. BlackRock’s consortium acquired Aligned Data Centers for $40 billion. These aren’t speculative bets - they’re strategic positioning for an AI-native future.

Building genuine momentum (the kind that produces compounding returns rather than one-off efficiency gains) requires four shifts in how AI engineering teams operate:

First, domain-based transformation rather than use-case proliferation. The firms reporting modest returns typically have dozens of isolated pilots. The firms achieving impact have restructured entire domains (investment research, client service, operations) around AI-augmented workflows. This requires harder organizational work but produces multiplicative rather than additive gains.
Second, governance as an accelerant, not a brake. The Investment Association reports 75% of COOs need clearer guidance on scaling AI from experimentation to enterprise execution. But leading firms have discovered that robust governance frameworks (clear accountability, transparent monitoring, compliant data handling) actually accelerate deployment by removing ambiguity and building stakeholder confidence.
Third, workforce preparation as a parallel track. EY’s research shows 97% of firms report minimal headcount changes currently, but 68% anticipate substantial workforce transformations within five years. Two Sigma’s Matt Greenwood captures the dynamic: “The future isn’t AI replacing humans. It’s humans who use AI well, replacing humans who don’t.” Teams that invest in AI literacy across roles move faster because they face less resistance and get more intelligent use-case identification.
Fourth, designing for the agentic future while delivering near-term value. The firms leading today are already architecting workflows where specialized AI agents can be introduced progressively. This means building modular systems, clear handoff protocols, and monitoring infrastructure that can accommodate increasing automation without wholesale redesign.

The patterns of success are now visible. Leading firms align AI strategy to business outcomes rather than technology capabilities. They invest in foundational infrastructure before scaling use cases. They build governance, oversight, and compliance from the start rather than retrofitting after incidents. They prepare their workforce and develop AI-literate talent across functions. And they design for an agentic future while extracting value from current-generation tools.

The task ahead isn’t adopting AI. It’s deciding whether your firm will be structured around AI as a competitive advantage or structured around legacy processes with AI bolted on. That’s not a technology decision. It’s a strategic choice about what kind of organization you’re building - and the 18-month window reflects how long you have before that choice gets made for you by market expectations you can no longer meet.