Successful AI implementation requires clear direction and well-defined outcomes from the outset. It depends on meaningful upfront investment in governance, planning, and a precise understanding of where AI delivers incremental value over existing automation.

Is your organization full of AI tourists or operators?

Innovation theater is a symptom. The underlying condition is an organization full of AI tourists. Tourists proliferate when investment precedes intention.

Most firms have now committed to a major enterprise AI platform. The question is whether that decision was anchored in a clear deployment plan or driven strictly by competitive pressure. Without that clarity, tech tourism can take over – experimenting, building, and launching pilots to justify the investment rather than create value.

Conversely, operators consciously think about how to embed AI into actual workflows where it actually changes outcomes. They establish frameworks that evolve with the technology and prioritize high-impact use cases. The result is buy-in: business users will now reach for the tools because they work.

Is your firm moving too slowly in AI adoption?

Asset managers are moving at different speeds in their AI journeys, and faster isn’t always better. The same Grant Thornton survey found that 34% of asset management leaders cite competitor activity as the primary driver of their AI investments. The fast movers often have greater financial flexibility and fewer data constraints. Rather than pursuing firmwide overhauls, they focus on targeted, high-impact use cases, which can be a pragmatic approach.

Learning from competitors, however, is not the same as reacting out of sheer FOMO. “Wait-and-see” firms can benefit by copying what works, allowing early adopters to absorb the initial costs and missteps. It’s a sensible approach for a technology whose innovation curve is almost completely vertical.ⁱⁱThese firms are typically more traditional, risk averse, and deliberate, with a strong emphasis on data protection and minimizing costly errors. Regardless of which firm you are, the objective should be consistent: aim to fail small and fast, test ideas quickly, and scale what proves effective.

Moving toward higher value AI use cases

Moving cautiously isn’t inherently a bad thing - it can reflect thoughtful decision-making, careful risk management, and a deliberate approach to progress rather than impulsive action. The firms moving fastest are prioritizing quick go-to-market use cases, such as investment screening, unstructured document reading, client reporting and onboarding, and internal knowledge sharing. While financial services sectors are leading in AI investment, they’re behind in other sectors in terms of “deep transformations” of core offerings. This could be traced to rigorous compliance concerns and data complexity. Most successful fast-movers are skimming the surface, implementing AI for quick-win workflows.

Across investment management, firms share a common ambition. They want to make information accessible through natural language. Instead of navigating fragmented systems and manual processes, they want to ask simple questions and receive immediate answers – whether that’s current returns in North American real estate, live exposure to private debt, or portfolio-level risk insights that previously took hours to assemble. With that said, AI is less about replacing humans, and more about accelerating their capabilities.

AI adoption in investment management relies on consolidated data

These agentic AI dreams rely on clean and consistent investment data across positions, cash, P&L attribution, margin, and attribution. Buy-side firms have only scratched the surface of AI use cases because the foundational operational data layer remains constrained by legacy infrastructure challenges. Thirty-one percent of asset management executives cite data quality and integration as a primary barrier to scaling AI, with data fragmentation and legacy processes hampering their efforts to change.ⁱⁱⁱ

Without a data layer that natively understands financial instruments and corporate actions, higher-value AI initiatives such as autonomous trade reconciliation or predictive risk management, carry meaningful risk: errors at this level can translate to significant financial losses or regulatory penalties.

The goal is a consolidated, centralized source of data – a golden copy that feeds upstream AI and operational systems. Firms are often operating on fragmented, legacy infrastructure, with multiple systems acquired over time, making data centralization both expensive and slow. A firm’s technology infrastructure must be able to model and normalize data from assorted disconnected or outdated systems, unstructured formats like PDFs and emails, and complex identifiers and classifications to ensure comparability. Standard LLMs tend to struggle to reason on large financial datasets without specific context, which is why we recommend investment-domain-aware AI models that understand financial instruments, corporate actions, and industry-specific matching logic.

Re-energize the firm to speed AI innovation

To reach the “dream state” of an AI agent that can provide real ROI, firms must first do the unglamorous work – the data equivalent of eating their vegetables: standardizing schemas across legacy systems, normalizing identifiers, and resolving material breaks before they reach the AI layer. Along the way, they can push back against AI fatigue – the skepticism and burnout that come from too much experimentation and too little payoff – by tightening success metrics and shifting from endless innovation theater to structured, outcome-driven implementation. In the end, the firms that scale AI successfully won’t be the ones chasing the flashiest model or picking the trendiest vendors, but the ones that quietly built the right foundation first.