One way in which advisors have found time savings using AI is in document extraction. Forms that used to have to be manually entered multiple times in different systems can now be scanned in mere seconds.
But there are still challenges. These tools can still make mistakes, and older systems often need to be modified to make the most of these new capabilities.
Those issues and more were the focus of the “From PDFs to Planning: How AI Is Transforming Document Extraction, Creation and Analysis” session last week at Financial Planning’s ADVISE AI conference in Las Vegas. The panel was moderated by Ian McKenna, founder and director of the Financial Technology Research Centre, and featured Grant Blindbury, partner at FMB Wealth Management, and Erica Molina, vice president of product management at Cambridge Investment Research.
Double-check everything
AI-powered document extraction technology has been a game-changer for both large firms like Cambridge Investment Research and smaller ones like FMB Wealth Management.
But despite the technological advancements, Blindbury said complex documents still present difficulties.
For example, some documents, like tax returns, are standardized. But others, like investment statements, can vary widely in formatting depending on the brokerage house, he said.
“We still have to keep an eye on those,” he said.
READ MORE: This is the biggest cybersecurity threat for wealth firms
In his practice, Blindbury uses FP Alpha for tax, estate and insurance planning. He said it is “excellent” at figuring out complex documents, pulling out handwritten notes and understanding the difference. That doesn’t mean the AI tools should be left to extract this sensitive data unsupervised.
“It’s getting better, but be careful,” he said.
Agentic AI frameworks help process documents
Many newer tech products were designed with AI built into them.
But not all software was created with AI in mind. For those legacy systems, AI overlays have become a viable option for firms working with older or more static technology.
Molina said her firm is working to eliminate duplicate, manual data entry that persists in its existing technology stack.
READ MORE: Using AI to write that client email? Think twice.
To address this, Cambridge Investment Research has been partnering with AI-powered wealth management platform ARQA to create an agentic framework overlay.
“We can’t fix all of the manual processes,” she said. “It would be great if we could start fresh and say, ‘All this new technology is going to work the way we want it.’ That’s just not realistic. We still have legacy technologies that we have to work around. And what we found is, by putting these agents on top, or having smaller agents that can do pieces of the work, we can put that operator on top of it.”
Molina said those in the back office of her firm refer to these AI agents as “digital associates.” They even have employee numbers and are supervised and managed similarly to their human counterparts.
“They are taking the part that our associates don’t need to be doing, which is taking a look at a data repository of a PDF version of accounts that have come in, reading it and then kicking off the process that we needed to do to, either open the account directly within our system and then act the direct business carrier, or through our multi-clearing firm,” she said. “We want to find the bottlenecks where people don’t do their best and highest work, and start to automate that and give it the tools, the process and connectivity to be able to have that agent kick it off for us.”
