Why AI-first thinking fails institutional operations teams

Technology leaders and asset managers warn that digital transformation without a clear business outcome leads to failed transformation and wasted capacity

Why AI-first thinking fails institutional operations teams

The first panel at CIBC Mellon's technology showcase opened on Wednesday with a startling message for institutional investors: don't start with technology.

The annual CIBC Mellon technology and innovation showcase, held at Oasis Immersion in downtown Montreal, brought together both fintech leaders and institutional investors for a day of “tech talks" along with a deliberate provocation from a group of asset managers and technology leaders who argue the financial services industry has the transformation equation backwards.

“Technology has never been the problem. It's been the abundance of it, it's been the application of it, and the governance of it,” said Santosh Joy, chief customer officer at ActiveOps in New York, adding how the real problem is that organizations have too many options and too little discipline around why they’re adopting any given one.

“It's going to be difficult, so when you apply and adopt tech, know first why you're adopting it. Then you've got to build the application, the orchestration, or the governance behind it. If not, you will be one year down the line still looking for value, cost, and impact,” he said, while also drawing a hard line between automation and transformation.

"Automating a task is not transforming a process or organization. Just because you fix the task doesn't mean it's a big thing in transformation," he said.

Decision intelligence over intuition

Sal Ammirato built on Joy’s point by targeting the language the industry uses. He argued that “digital transformation” as a label has become misleading because it puts the emphasis on the technology rather than on the business problem it’s supposed to solve.

“We constantly hear the word digital transformation, but the reality is that it’s not going to drive a better outcome for an organization. It’s really what is the business transformation that is then enabled by technology,” said Ammirato, Toronto-based VP and head of retail distribution and channel strategy at Sun Life Global Investments.

James Maxfield, chief client officer at England-based Duco, described a fundamental shift in what operations teams should be doing. Resolving exceptions after the fact, he argued, is becoming the minimum expectation rather than a differentiator. The real opportunity, he suggests, lies in moving from reactive cleanup to proactive pattern recognition, where firms can tell clients about risks before they cause damage.

He illustrated the point with an example from agency lending, where a firm could trace a string of operational risk events back to a single root cause and flag it for the client pre-emptively.

“Being able to start to go to your customers with those types of insights, I think, will be ultimately where the value comes from,” he said.

The cost of deploying AI too early

While Maxfield acknowledged that senior leadership is drawn to AI-first thinking, particularly around agentic technology, he suggests that can create real problems when experimentation moves into production without controls.

“When that moves into production in an uncontrolled manner, which does happen, that’s not necessarily a great outcome,” he said, adding that the industry’s current architecture challenges are largely the result of the same undisciplined pattern repeating over two decades.

The tendency to start with technology comes as financial services firms are under growing pressure to rethink how they operate, while investors want information instantly and advisors expect seamless platforms, said Ammirato. As a result, asset managers face the consequences when the back office can't keep up.

"Investors today are continuing to look for information that is at their fingertips, that is readily available when they want it available,” he said. “Advisors have a lot of choice on who they want to partner with."

The stakes, he argued, aren’t just operational, particularly when it begins to cause friction.

"It becomes a commercial issue, and unfortunately, it's tough to overcome at certain points," he said.

Measuring what transformation actually delivers

To that end, Joy argued that most firms can identify what’s frustrating them but have little visibility into where their capacity goes, highlighting that standard SLAs (service-level agreements) and KPIs (key performance indicators) don’t capture it.

By combining basic operations data and applying decision intelligence, he believes organizations can build an evidence-based picture that separates genuine opportunities from surface-level irritants.

“What most organizations don’t know is how capacity is being consumed,” he said.

He pointed to CIBC Mellon’s fund accounting operation as a case in point, where the approach surfaced more than 10,000 hours being spent on reconciliations. That data allowed the firm to diagnose the root cause, redesign the process, and only then bring in automation through Duco. The sequence, he stressed, is deliberate and cyclical.

“You find what the problem is, you fix the process, you apply the technology, and then you go back and repeat it over a consistent period of time,” he said. “That approach will help identify what's a potential opportunity and what's essentially causing frustration and pain.”

Business transformation, not digital

Ultimately, Ammirato stressed that technology transformation - and the industry - needs to start with the outcome the end client is trying to reach - whether that’s an advisor, investor, or dealer. Without that firmly in place, firms risk pouring resources into whatever problem happens to be most visible, which may have little to do with what actually drives the business.

In asset management distribution, that means retention targets, advisor experience, and investor experience. If the operational fix doesn’t connect back to those drivers, Ammirato emphasized it doesn’t matter how urgent the issue appears.

“If we are solving for a visible issue, one that’s presented in front of us, it might not actually be the right challenge that we actually should be addressing,” he said. “Just because it’s visible doesn’t mean that it aligns to the business growth drivers of the organization.”