Advisors racing to add on new platforms may find their legacy plumbing can’t support the next wave, warns one expert.
Wealth management firms have raced to bolt artificial intelligence onto their technology stacks this year, from meeting note takers to pilot projects promising new levels of automation. But behind the widely reported boom in productivity, many are now discovering just how unprepared their data plumbing is for a true AI-driven future.
That’s according Garrett Oakley, partner at consultancy Alpha FMC, who has spent his career across advisory, fintech and consulting roles in the RIA and independent broker-dealer channels.
“Wealth management, at the end of the day, is professional services,” he said in a recent conversation with InvestmentNews. “Every firm is trying to figure out how you become more efficient with your most expensive resource, which is people.”
‘It saves time, but that’s where it stops’
For many firms, that has translated into rapid adoption of AI notetakers and similar tools, which promise quick productivity gains without forcing firms to fully rewire legacy workflows or clean up old databases. In Oakley’s view, that is precisely why they became the first widespread use case.
“Notetakers were the easiest way of dipping your toes into it,” he said, pointing to how they can “take in‑the‑moment data and leverage it” straight from client meetings.
But that convenience can only go so far. According to Oakley, meeting transcripts and summaries often land in a CRM or shared drive, where they add to an ever-growing mass of data silos. Advisors might save time on manual note entry, but the insights from those conversations often go nowhere.
“It saves time, but that’s where it stops,” he said.
The “next big thing,” in his view, is agentic AI that can interpret those notes and kick off actual follow-up work — drafting emails, updating plans, flagging tasks, or pushing data into other systems — without waiting for the advisor to drive and steer each step. But most wealth firms still don’t have the structured, connected data layer that those agents would need.
“Everything comes down to data,” Oakley said. “What AI is starting to expose in these early days is that the plumbing is not there.”
The high cost of legacy tech
The fundamental issue, he argued, is that many mid-size RIAs and independent firms still operate with a loosely connected stack of tools that were never designed with a centralized data strategy in mind.
Part of the issue is historical. For a long time, data was not treated as something advisors or smaller firms needed to actively own and manage. Tech stacks were typically built around several pillar vendors — a CRM, a planning platform, and a portfolio system — with point-to-point integrations considered “good enough.”
“For the longest time, no one really asked where all that data goes,” Oakley said.
But as firms push deeper into AI, he expects more leaders will reckon with the cost of that legacy sprawl. Beyond the initial “glitz and glamour,” the real question becomes whether the infrastructure can support enterprise AI across hundreds or thousands of advisors.
“Firms are saying, foundationally, we are not ready,” Oakley said, describing clients that raced to showcase proofs of concept but are now hesitating on rollouts across a national field force with complex supervision, compliance, and data-governance requirements.
As firms grow and start to see the limits of fragmented integrations, he believes vendors will increasingly need to be “integration-agnostic” and support open architectures, allowing firms to route data where it needs to go. Those that refuse to connect to the broader ecosystem are the ones he sees at risk.
As for firms trying to position themselves for the next wave of AI, Oakley’s advice is to start not with a tool, but with a plan.
“It’s amazing to me how many firms just haven’t even had or don’t have an ongoing view of their own data strategy,” he said. “Step one is: what are we trying to accomplish, and how does our firm feel about the best path for us? Then you get to step two, which is the solve. That might include using agentic AI to normalize, clean, and maintain data.”
From there, firms can move on to deploying the technology. On that note, he said firms should fight the temptation to keep growing their tech stacks, without considering where and how to prune.
“We continuously talk about these new technologies, but we never talk about retiring anything,” he said, pointing to an “everything plus” mentality that “makes the data game even more complicated.”
Looking ahead, Oakley thinks the industry is still in the early innings of AI. While many firms are framing use cases in terms of their current processes, he argues the next stage of innovation will require them to ask how those processes themselves might change.
“It reminds me of the Henry Ford quote: he said if he had asked people what they wanted, they’d have said faster horses. We’re doing something similar with AI,” Oakley said. “The real opportunity is to ask: what does the future look like outside the confines of how we operate today?”
