'They're going to use their business sense. It's not just the science that AI would give them,' says CIBC Mellon's head of enterprise security.
While AI promises to boost efficiency and provide new insights across all industries, investment managers aren’t rushing to hand over their decision-making power just yet.
Rather, they’re exploring where AI fits and where it doesn’t within their operations, explained Mike Plantinga.
“When I talk to chief investment officers, they tell me they’re not using it as part of their decision-making process when they’re trying to figure out where they’re investing,” he said.
As vice president and head of enterprise security at CIBC Mellon, Plantinga has regular conversations with institutional investors and how they’re utilizing AI.
“They’ve been doing this for 30 years, and they’re going to use their business sense. It’s not just the science that AI would give them. It’s the art they layer on top that's going to help them with the information to make the decisions on how the portfolio is going to be balanced and where they're going to really move their investments to,” he explained.
While AI can aggregate data and enhance predictive analytics, most investors are still skeptical about letting it drive portfolio decisions.
“AI may help them aggregate data better, faster, stronger, but it’s not necessarily being truly embraced for the capabilities of AI,” he said. Instead, firms are sticking to the tools they already trust, such as predictive analytics models that have been in place for decades.
Where AI is making an impact, however, is in back-office operations. From automating administrative tasks to improving meeting documentation and report generation, AI is streamlining processes that once took up valuable time.
“Can it do a transcription? Sure. Can it give me some narrative to help me with portfolio aggregation? Sure, but it hasn’t really taken shape at the transaction level yet,” asserted Plantinga.
Even on the operational side, institutional investors are still searching for the “killer use case” that will fully justify AI’s integration into their workflows.
“They’re no different than the rest of us,” Plantinga said. “We’re all trying to figure out what’s that killer use case we’re going to use in the business for the delivery of AI and its real value add for us.”
One of the biggest roadblocks is data readiness because it requires structured, accessible data to be effective. This is what most organizations are still trying to figure out.
“You take AI and install it somewhere in your tech stack, great. Now you have a technology that has no access to any information. How do I take this great tech that has a lot of capabilities and now give it accessibility to the information that's going to give me an outcome that I'm desiring?”
Beyond data challenges, firms are also grappling with a skills gap as AI requires a different level of expertise; not in coding, but in knowing how to ask the right questions.
“The prompts in AI, that’s the equivalent of writing a piece of code,” explained Plantinga. “Before, I would write a SQL query that says, ‘Select all the stuff for my product and match it to my transactions.’ Now, I need to use natural language to ask AI to do that.”
Employees also need to be trained not just on how to use AI, but on how to validate its outputs.
“It may be accurate, but does it have quality? If the data set it’s using is incomplete, I’m going to get a false report,” noted Plantinga. “We’ve always called it garbage in, garbage out. AI is no different. If I ask the wrong question, I’m for sure going to get the wrong answer.”
For all its potential, AI also comes with risks. Plantinga sees three key areas where firms need to stay vigilant: client data protection, vendor AI adoption, and employee oversight as he admits he doesn’t always know when a vendor could wind up introducing AI capabilities into their tech stack.
Despite concerns about job displacement, employees can rest assured as Plantinga isn’t seeing AI replace human workers. He noted while there’s a lot of AI efficiency, that efficiency doesn't equate to job erosion.
Instead, it’s eliminating low-value tasks so employees can focus on higher-priority work.
“There’s a lot of fear about AI and job erosion, but so far, we’ve found no support for any of that thinking,” he said.
“It’s taking away those lower-value tasks that we expect people to be doing and allowing them to focus on higher-value tasks. That’s what people want in the first place.”
