The Impact of AI's Evolution on Asset Management

AI Adoption: Rapid Spread, Gap in Results

The evolution of AI is bringing about a structural shift in the asset management industry. According to a survey of 200 major global asset managers, as of the end of 2024(*1) only about 10% of firms were actively exploring and implementing AI tools, while the remaining 75% recognized its potential value but were still figuring out how to use it. Just one year later, by the end of 2025(*2) , 70% of buy-side firms had moved to deploying AI in live front-office operations. However, only 17% of firms have reached a "mature" stage, i.e. deploying AI across multiple business areas with measurable ROI, meaning a substantial gap remains between "adoption" and "results." The industry as a whole has begun to embrace AI in earnest, even as its real effectiveness is only now being tested.

In Japan too, a survey conducted by the Financial Services Agency (FSA) in October–November 2024 found that only around 10% of firms were using generative AI to enhance asset management(*3). This figure is close to the global figure from the same period (end of 2024), suggesting Japan's market was at a similarly early stage as the rest of the world. While no updated survey exists yet, given the global figure's rapid rise to 70% by the end of 2025, it is plausible that a similar acceleration is occurring in Japan as well.

Lessons from the Past: How the Internet Shifted the Source of Differentiation

The earlier experience of internet adoption offers a useful reference point. Before the internet became widespread, obtaining financial information itself carried a cost.

In the U.S., until electronic disclosure via EDGAR became mandatory with the SEC in the mid-1990s, the only ways to access corporate disclosures were subscribing to commercial data feeds or physically visiting an SEC reference room, meaning only the largest institutional investors had timely access(*4). The same was true in Japan, where securities reports were purchased as printed booklets and data was entered manually; the number of companies that could be analyzed was inherently limited by shelf space and staffing resources. Whether or not to analyze a company was not just a matter of capability, but of whether resources could be devoted to it.

As the internet spread and electronic disclosure infrastructure matured, the cost of obtaining information fell sharply, and asset managers large and small gained access to the same information at the same time as their competitors. As the information advantage disappeared, the source of differentiation shifted from "obtaining information" to "how quickly and deeply that information could be processed and interpreted." AI may now bring about a similar leveling in this "processing and interpretation" stage.

Impact on the Investment Workflow: Automation of the Middle Steps, and the Growing Importance of Both Ends

The asset management workflow begins with establishing an investment philosophy, moves through middle steps such as building investment hypotheses, research, analysis, investment decisions, and portfolio management, and culminates in engagement with portfolio companies. Of these, the middle steps are relatively standardized and represent an area with particularly large scope for AI-driven efficiency and automation.

According to McKinsey estimates, high-impact use cases such as AI co-pilots for portfolio managers could deliver productivity gains of 25–40%(*5). As automation of the middle steps advances, the areas where humans should add the most value are likely to converge at both ends; upstream, in building investment philosophy, and downstream, in deep engagement with companies. This resembles the "smile curve" structure often seen in manufacturing industry analysis.

A Caution: The Gap Between Adoption and Results

At the same time, we should be careful not to be overly optimistic.

According to McKinsey, asset manager margins have fallen by 3 percentage points in North America and 5 percentage points in Europe over the past five years, even as technology investment has continued to rise — a rise that has not necessarily translated into productivity gains*5. Simply adopting AI is not enough; real value emerges only when it is combined with a redesign of the workflow itself. As the source of differentiation shifts, what will be required of asset managers is to remain open to change, keep experimenting, and continue asking what constitutes genuine added value.

Closing Remarks

At this juncture, what is needed is for the diverse players across the investment chain, including asset managers, corporate IR, institutional investors, and others, is to pool their knowledge and explore how to build a better ecosystem by harnessing technological evolution. We too aim to remain open to change, keep experimenting, and continue improving the quality of our judgment as long-term investors.

(*1) SimCorp, "Perspectives from 200 Global Buy-Side Operations Leaders"

https://www2.simcorp.com/investops2025

(*2) SimCorp, "The AI race in 2026: Building foundations for a lasting advantage"

https://www2.simcorp.com/investops2026

(*3) Financial Services Agency of Japan, "AI Discussion Paper (Version 1.1)" (March 2026)

https://www.fsa.go.jp/news/r7/sonota/20260303/aidp_version1.1.pdf

(*4) Chang, Ljungqvist, and Tseng, "Do Corporate Disclosures Constrain Strategic Analyst Behavior?" (The Review of Financial Studies, 2023)

https://ssrn.com/abstract=3579466

(*5) McKinsey & Company, "How AI could reshape the economics of the asset management industry" (July 2025)

https://www.mckinsey.com/industries/financial-services/our-insights/how-ai-could-reshape-the-economics-of-the-asset-management-industry