Brown Brothers Harriman has been busy—so busy that in the past two years, the bank’s investor services unit has delivered three large-scale tech platforms underpinned by various forms of artificial intelligence, and is set to roll out a fourth to its internal users in July.

The most recently implemented project, Guardrail, was completed in 2020 and is part of an ongoing endeavor to radically transform its fund accounting process. It expanded upon one of BBH’s first AI projects in the division—the Anomaly NAV Tracking System (Ants), which was developed using supervised machine learning to help BBH’s analysts weed out thousands of false anomalies during end-of-day net-asset-value (NAV) strikes.

The Guardrail application monitors those NAV movements against a benchmark and historical performance and flags variances. It goes a step further than Ants does by identifying possible reasons for abnormal deviations, and brings those automated insights straight to BBH’s analysts through natural language generation—a rising priority for the bank, which is using the emerging tech to generate “smart alerts” that act as idea generation for research projects, Kevin Welch, managing director of investor services at BBH, tells WatersTechnology.  

“We did what we said we were going to do, and it’s in place. Now what we’re doing is, let’s say security pricing makes up the vast majority of reasons why we see these deviations at the fund level to benchmark—that’s already automated. We’re looking at the next level down of reasons, and we’re automating that,” he says. “So what you’ll get at the final view of this is your deviation from the benchmark, and you’ll be able to say security pricing looks good, but then look at corporate actions and be able to go through what used to be four or five separate checks and reconciliations.”

The forthcoming AI-based algo validation tool is meant to help BBH’s own business users more easily verify their own proofs of concept (POCs), taking some of the onus off its data scientists and IT resources, who have often been caught in the middle of explaining results or finding the necessary data. The application was already in use by the bank’s clients, but after positive feedback, Welch and others in his business unit decided to deploy it throughout the organization, beginning next month.

The thesis of this tool is that if a business user identifies a problem that technology can potentially solve, they should be able to quickly access and scale the model, test their theory in real time with real data, and interpret the results for themselves. For example, if a BBH user wants to experiment with predictive analytics for forecasting trade fails, the platform will allow them to locate the data they need in the bank’s system, onboard it, normalize it, and then view the results dashboard format. From there, they can choose to tweak the datasets or try out different models that have been developed by IT.

“One of our lessons learned is that AI is really powerful and a lot of times the allure of demonstrating the raw capabilities of AI can outshine delivering actual business value. So we’ve been really focused on where we can use AI to deliver lasting value. And more and more as we thought about this, we realized it’s hard to do that in a lab or IT group that’s really separated from the business,” Welch says.

Sitting alongside Guardrail, Ants, and the algo validation tool in the investor services business is Linc, the foundation upon which the others were ideated and built. Linc is an artificially intelligent reconciliations tool that reviews positions in BBH’s custody and accounting records and automatically routes and matches them. Though it’s well into production and not undergoing significant updating, Welch says the bank is getting more than 95% matching accuracy through Linc now.

Welch says BBH has no plans to commercialize any of its AI products at this time, though it is a prospect he expects the bank will likely consider as reconciliations, striking NAVs, anomaly detection, and POC validation are wider industry problems.