Asset managers aren’t investing as aggressively in artificial intelligence as the hype might suggest.
Despite the talk surrounding AI’s ubiquity and inevitability, new research from the Thinking Ahead Institute and CAIA Association projects that asset managers expect labor costs for customer facing employees to remain the same over the next five years, while only modestly raising their technology spend, from 12 percent now to 14 percent five years from now (this includes total spending on technology — hardware, software licenses, data vendors, IT personnel — and not just AI).
"It seems like most asset managers are still in the experimentation phase with AI, but it's not a reflection on the technology,” Aaron Filbeck, a managing director at CAIA and co-author of the report, told Institutional Investor. “The AI use cases are still being tested, so despite the hype and interest in AI, I’m not all that surprised that the capex being spent on these tools is not moving as quickly as the headlines might suggest.”
Filbeck noted that while the use cases for AI “keep getting better,” managers are still using the tools selectively to improve research, not to overhaul the company’s structure. “You hear about research teams utilizing the tech for a research project to speed up or widen the data set so that’s where we’re currently at in terms of adoption of AI,” he said, adding that even though there are “lots of headlines about back-office capabilities, we’re still very early in terms of adoption.”
For established companies, the real constraints are fragmented data and incomplete infrastructure. Additionally, governance frameworks for evaluating and approving AI use cases are still being built. Compliance and risk teams are also learning how to incorporate AI into their practices.
Still, allocators are being encouraged to adopt AI.
Shaun Wei Tjia Ng, a former allocator with Cleveland Clinic who authors the educational AI for Allocators newsletter, told Institutional Investor that most allocators treat AI as an IT problem, when it is instead “a change management and leadership problem.”
Ng explained that CIOs “have to be much more hands on” with this tech. “They need to be a constant involved sponsor that asks questions, that tries things themselves, that tells their team members, ‘I tried or failed at this.’” Allocators also need to have a proactive vision for AI in their investment office and not just reactively acquire the latest tech.
Given the rapidity of advances in AI and the heavy information content of asset management, Paul O’Brien, trustee and investment committee member at the $13 billion Wyoming Retirement System, is skeptical of any forecasts about five years from now.
“Asset managers and owners are inherently cautious. But the potential disruptions ahead are enormous,” O’Brien said over email. “I would not encourage that caution.” Instead, he advises making understanding AI a “board-level priority and make training in AI mandatory.”
