Over the last few years, I’ve heard from a handful of advisory firms taking a technology approach different from most.
These firms, often founded or led by youngish-yet-veteran advisors, and usually tech-savvy and with engineering connections, start out unsatisfied with off-the-shelf solutions to their technology needs. They, in turn, or from the outset, build some or much of what they need themselves. Most do not reinvent the wheel and build everything in-house; instead, they focus on core components.
I’ve recently spent some time with the founders and some advisors from Compound Planning.
Alex Farman-Farmaian, one of its co-founders and co-CEOs, said the firm’s ethos remains the same as I’d previously heard a couple of years before, that of an RIA serving a range of clients from mass affluent to high-net-worth, functioning like a digital family office.
While part of what is Compound launched five years ago, the firm fully came together in its current form in 2023. That’s when co-founder Christian Haigh’s two-year-old RIA platform, Alternativ Wealth, which had been working to build back and middle-office software to help advisors manage their books, bought Compound, a tech-enabled financial and tax advisory for tech executives.
Compound now manages more than $4 billion in client assets (it crossed the $3 billion threshold in January, so growing at a pretty fair clip) and now has 50 advisors spread out across the country.
Tara Shulman, a principal financial advisor at Compound, who works out of the firm’s New York office, has been there since near the beginning. She began her career immediately after college with UBS, spending five years there before landing at Janney Montgomery Scott for another 3 1/2 years.
“Those are big ships that are very hard to turn and are running on very old technology,” Shulman said.
“I’ve written a lot of tickets about trading being down,” she said, recalling having to use the old vacuum tube system to send trades to traders on another floor on occasion prior to her time at Compound.
An IronMan Suit for Advisors
Shulman said that Compound’s clients demand and expect a different experience, with many being technology founders or executives themselves. For example, advisors at Compound rely on Activity Monitor, a highly configurable dashboard that provides constant real-time data on client accounts and portfolios and generates proactive insights on clients being pulled into Advisor HQ, the firm’s core AI-driven CRM platform, both of which were built entirely in-house.
CRM was one of the core components the firm decided to own from the beginning, according to Nelson Arnous, who heads up product development at Compound.
“We are building an Ironman suit for financial advisors, CRM maybe can be thought of like the orb in the middle [of Ironman’s chest],” Arnouse said in an earlier interview, where he also elaborated on the firm’s use of OpenAI’s ChatGPT 4 as its foundation large learning model.
While ChatGPT 5 has been released, the Compound engineering team, comprising 10 engineers and developers, has been waiting to evaluate its performance and consistency before considering a switch to the newer model.
As someone who writes specifically about technology, it has become time for me, as much as it might feel like delving into the weeds, to begin going below the surface of simply stating that a particular firm or project is using a particular large language model, and in this case Compound’s team is using Langfuse, which is a collaborative platform for prompt management. It enables the team to manage testing and versioning, and deployment of prompts without requiring code changes.
While OpenAI was initially selected over other foundation models, Compound’s developers still plan to use different providers where they’re best suited to the use case.
Use Cases
While Compound is building core components of its technology stack, which already includes onboarding, account opening and a proposal generator, it sometimes integrates third-party tools where engineering the underlying technology doesn’t make sense.
One such example is the firm’s launch of tax filing earlier this year. The firm did so in partnership with the startup april, a tax forecasting, planning and data platform that they embedded into their own technology (april has now raised more than $78 million in funding after a Series B raise in July).
Farman-Farmaian provided another example of the firm’s approach.
“Portfolio accounting, we don’t plan to be in that business, so we partner with Orion there, and they have an API we can hit,” he said, explaining that because of the application programming interface, the advisor does not have to log in to Orion itself and can remain in their day-to-day Compound experience and environment.
Another of Compound’s advisors, Jonny Jonson, who is based in Washington state, has now been with the firm for 18 months.
“I started my career at a small RIA, and on my first day, the printer broke,” he said, noting that he had to fix the problem. When he started at that firm back in 2017, it was using Salesforce for CRM, MoneyGuidePro for financial planning and SS&C Black Diamond for portfolio reporting.
“When I had to prepare for a client meeting, I was going all over the map to prepare and then myself and an associate advisor would be taking notes [during meetings] and at the end of the day we would have a pile of notes and to dos,” Jonson said, noting that at best any assistance with integrating technology either had to come from the providers themselves or required the hiring of a third-party consultant.
“At Compound, having a dev team right here that can iterate immediately on something is light years ahead,” he said.
“We’ve figured out how to plug AI into our Advisor HQ and really enhance it, giving us the ability to generate service tickets right now, where I can say,‘Hey, we have new accounts, can you go ahead and copy over these beneficiary designations from these previous accounts?’” Jonson said, going on to describe the tedious and manual process that had been involved in doing the same thing at his previous firm (and that likely remains as tedious today for many advisors).
Another example that he said was related to this and is a next step, is incorporating AI into recording and analyzing client conversations, something that some third-party AI-driven notetaking platforms including, GReminders, Jump, Zeplyn and Zocks are known for.
Jonson described, for example, the frequent conversations he and other advisors have regarding taxes. Soon, Compound’s AI will be able to search through every advisor-client interaction, whether notes, emails or transcribed meetings and calls, and automatically deliver back to the advisor a breakdown and summary of every time taxes came up and where the client already stands on them.
“That way, when we’re having that November tax planning meeting with clients, I do not have to go through Salesforce [he used this as an example], work through the emails and search everything I have on my end, we will instead have a nice little summary generated, which is honestly a dream come true,” he said.
What’s Next
Compound’s head of product development, Arnous, said that as an initial example of an AI agent embedded within the firm’s technology platform, a service request agent was currently in early testing with a small group of advisors.
“Basically, as an advisor or associate, you can ask the agent to request a trade, and it will structure the ticket and hand it off to the custodian operations team.
“At the end of the day, we are building a general assistant for the advisors, an AI assistant,” said Arnous, one which will end up with many capabilities.
This is the first in a series of columns that will discuss the varying approaches RIA firms and some of the most innovative third-party providers are taking to weave advanced analytics and AI capabilities deeply into their technology stacks.