Why data, not AI, is the foundation for the next stage in wealth tech

In the rapidly evolving world of wealth management, artificial intelligence is quickly coming up as a force multiplier that will redefine client service, investment research, and operational efficiency.

But as Chris Edwards, director of solutions at SEI cautions, the real magic of AI can only happen if the industry gets its “house” in order first.

“Data really is the foundation ... AI becomes the doors and the windows and all of the pieces inside the home that make it special, unique, and different," Edwards told InvestmentNews in a recent interview. "That allows you to really modernize, make it beautiful. But building anything on a shaky foundation means that inevitably there’s an opportunity for it to fall.”

The hidden debt of tech sprawl

For many firms, that shaky foundation is the result of years – sometimes decades – of technology “sprawl.” Jeff Benfield, head of Advisor Solutions at SEI, describes it as a proliferation of legacy systems, each built independently and rarely designed to work together.

“It’s not necessarily an issue with a legacy system. It’s the fact that I [as an advisor] have seven legacy systems that I use, not one,” Benfield explains. “Sprawl in and of itself has created the lack of normalization.”

Apart from the patchwork of platforms, which was made worse by the rush towards digitization during the COVID-19 pandemic, the lack of consistent processes has also made it difficult for firms to standardize data and workflows.

“Processes haven’t necessarily caught up with the modernization of the technology to make sure that there is standardization within those in order to allow the technology tools' capabilities to keep up,” Benfield says.

Private markets: A wild west of data

If public markets are the well-ordered rooms of the house, private markets are still under construction. Edwards points out that while public markets today benefit from standardized identifiers and transaction types codified over the course of decades, “a lot of the data [in private markets] is still very unstructured – creating a potential speed bump for the so-called retail revolution.

“When we talk about just the simple consideration of the subscription document – what does it take for [a firm's] investors to enter a fund? – No subscription document looks like any other one,” Edwards says.

With a dizzying array of partnerships developing in real time between private market players, traditional asset managers, and wealth tech platforms, Edwards sees a complex landscape of data streaming in as PDFs or even “wet ink” forms, each requiring manual processing and interpretation.

“Automatically, you’ve got this very different environment of data of how it’s being handled. So the cleanliness or the completeness of that data will vary,” Edwards says.

Standardization: The next frontier

As more private market products like semi-liquid funds and private market ETFs come online, the problem of standardization is moving slowly but surely from strategic importance to competitive urgency.

“Once you make that barrier to entry lower and those account minimums go way down, the standardization is going up. It has to, because of the volume [of transactions],” Edwards notes. “Your data platform [has to be] ready for that volume when it comes time to interact with a retail-type investor or retail type of platform.”

Benfield agrees, emphasizing that the integration layer – the connective tissue between systems and partners – will be the likely target for enhancement.

“The easier you can do that, whether that’s just to pass data back and forth or to have AI models work off of, I would say the integration layer is probably the likely candidate [as a focus of enhancement efforts].”

Attention advisors: Don't go it alone

At this point, wealth firms applying AI across channels have notched countless easy wins from applications like note-taking and investment research. But as advisors search for the next killer application for AI, Edwards says firms must take a hard look at the structures and processes supporting their data.

“Data is going to be extremely important to the success of the long-term vision of AI, even if there’s short-term value that’s being felt today,” Edwards says. “Those that get the data infrastructure and the data models correct will have much more value in the future when it comes to the utilization of AI.”

Given the continued evolution of AI models with updates emerging every few months, Benfield says it raises more even questions for advisors – knowing which LLM to use, how to integrate built-in AI services into their daily worklows, or how it can extend into the models they provide for clients. 

“RIAs or independent advisors have been asked to be not only advisors, but CTOs ... They’ve been asked to do much more than the original job, which is why having a partner becomes so crucial," he says. "It’s partners who should be allocating capital and being really the representation of the CTO rather than asking the advisor to be the CTO, simply because where their value lies is in the relationships and the clients they serve.”