Family Office Investing Requires Mounds of Paperwork. AI Is Going to Help

AI is “going to be incredibly helpful in general but definitely within the family office and ultra-high-net-worth space,” says Mark Rogozinksi, head of family office services at Cresset in Chicago. 

Chief among these mundane tasks is the mountain of paperwork associated with alternative investments, such as real estate, private equity, and venture capital. 

One of these tasks is a requirement for family offices to fill out Internal Revenue Service Schedule K-1s for every partnership they are in, for example, Rogozinkski says. 

These partnerships “could be a private equity fund, a hedge fund; it could be two family members who buy a piece of real estate and stick it into a partnership for tax and risk-mitigation reasons,” he says. It could also be a family limited partnership with many partners, each of whom are required to fill out a K-1, disclosing their share of income, deductions, credits, and more, according to the IRS. 

“A family could have a tax return that’s 1,000 pages long because of all the alternative investments and additional reporting requirements,” he says. 

With AI, much of the data required by these filings could be collected and entered automatically into a tax reporting and compliance system, for instance. These kinds of systems are going to move from manual, to semi-manual, to completely automated in the future. “And that future is coming quick,” Rogozinski says. “It’s not 10 years from now, it’s a year or two from now.”

Finding a way to streamline manual processes that today involve pouring over emails, filings, PDFs, and the like, has been “the holy grail” for institutions for the past 30 to 40 years, he says. Automating these processes will have “a significant impact on the efficiencies and back office for not only family offices but anybody that receives a K-1.” 

More Time for Analysis

Rogozinski doesn’t expect the advent of AI within offices to mean fewer staffers are needed. Instead, the staff will have more time to spend on meaningful work. 

Currently, offices can spend 80% of their time gathering data, leaving only 20% of their time to analyze and understand it. 

With AI, only 20% of staff time will be required to review the data collected automatically to ensure it’s accurate, and then “we can spend 80% of our time analyzing it, coming up with better tax strategies, and ensuring that the return is correct,” Rogozinski says. 

In addition to tax forms, AI can also collect statements from investment custodians and outside portfolio managers, putting that data into a reporting system “and then enhancing that reporting capability through either connections to current systems that they have or even going out to a data set on the internet and collecting real lifetime data and extracting that and putting it in,” he says. 

Today, a lot of pricing or performance information on investments comes from disparate sources, and isn’t always timely. As a result, a family office’s reporting systems could include current data sitting alongside data that’s a year old. “AI could cull the internet or a dataset real-time, all the time, and when news hits, you know immediately about what’s going on,” he says. 

A family office CIO couldn’t necessarily react to critical news affecting a long-term real estate investment, for instance, but “that information could impact other things that you do from an investment perspective or from a cash flow perspective or a risk mitigation perspective,” he says. “More information in the hands of the decision makers will help them make better decisions.” 

AI should also make it easier for family offices to coordinate the various systems they use to track information such as portfolio performance, estate planning, insurance, custodial services, and investment rebalancing and trading. 

“Most of those systems are somewhat disparate,” Rogozinski says. Even when systems work together, it’s imperfect. The expectation is that AI systems will one day exist to “sit on top” of a data warehouse of reporting and investment information—systems such as Microsoft SQL Server or

Snowflake

—enabling a family office to extract relevant data and information and report on it “efficiently and effectively and consistently,” he says. 

In the next several years, the power of machine learning and artificial intelligence will be focused on automating manual, repetitive tasks, Rogozinski says. “The future past that is, ‘can AI and machine learning actually start helping you make decisions on that data set?’” 

AI, for instance, could help a CIO understand market trends, or assess the likelihood of future Federal Reserve actions on interest rates and the investment opportunities those actions could create. 

“It’s going to say, ‘sell Coke and buy Pepsi and here’s why, based on our projections and the data we can see,’” he says. “A human needs to review it, there will be discussions around it, but the evolution of AI over the next several years is going to get there.”