Form ADV Review Finds AI Still A Back-Office Story At Most RIAs

Registered investment advisors that told the SEC they use artificial intelligence added staff over the past year rather than cutting them, though nearly all of that growth landed in operations roles rather than among advisors themselves, according to a new report.

The report also found that only a small minority of firms were reporting AI usage in their brochure filings.

“Unlike previous technology waves in wealth management—portfolio accounting, re-balancing software, digital onboarding—AI has arrived without a settled vocabulary, an agreed-upon set of use cases, or a consensus on what advisers owe clients in terms of explanation aside from the regulatory requirement to represent AI use accurately,” stated the report, published by wealth-tech firm Astraeus in partnership with Pirker Partners.

Providing a window into how firms are thinking about AI, the report continued, the Form ADV Part 2A “brochure” filing requires advisors to describe their services, business practices, investment risks, methods of analysis, fees, and conflicts of interest in narrative form. This is filed annually and updated with material changes.

Just 370 firms, or 6% of the group, mentioned AI, machine learning or algorithmic tools anywhere in their brochures, representing 11% of industry assets under management, the report found. Adoption skewed heavily toward larger firms—16% of RIAs managing $5 billion to $25 billion disclosed AI use, compared with 5% of firms managing less than $1 billion.

Among the very largest RIAs, those above $25 billion in assets under management, only three of 33 firms in the sample disclosed AI use at all, a group too small to draw firm conclusions from, the report said.

“It is the only place in the wealth management industry where thousands of firms describe their own operations, in their own words, on a common schedule, under a legal obligation to be accurate,” the report said. “It is therefore an imperfect but uniquely comparable window into how the industry is thinking about AI.”

The report reviewed Form ADV Part 2A brochures filed by 6,384 independent private wealth RIAs as of March, alongside Part 1A business data from April 2025 and April 2026. The firms in the analysis were fee-only or hybrid RIAs with a substantial base of high-net-worth or individual clients.

There is no Securities and Exchange Commission requirement that registered RIAs disclose their AI use. However, firms that acknowledge this modernization to their clients include that information in their brochure, providing insight into other aspects of their business, such as headcount.

Firms that disclosed AI use grew total headcount by a median of 15% between April 2025 and April 2026, compared with 8% of firms that made no mention of the technology, the report found. But the composition of that growth cuts against a straightforward “AI is adding advisory jobs” reading.

At large and enterprise RIAs disclosing AI use, non-advisory staff grew by a median of 14.2%, more than double the 6.7% growth in advisor headcount at the same firms. Large RIAs that didn’t disclose AI use actually grew their advisor ranks slightly faster, at 7.1%.

“Large AI adopters grew median assets under management per advisor by 22%, versus 12% for non-adopters,” the report said. “The evidence does not appear to support the claim that AI will result in job eliminations in wealth management.”

The report admitted that there were limits on what findings its data could support. Firms that disclosed AI use were already growing faster than their peers before AI adoption picked up, the report said. For example, these firms posted a median four-year AUM growth rate of 11% versus 9% for non-disclosers, a gap that predated the current wave of AI investment.

Advisors reading headlines about the report should also be careful about what “disclosure” means here. Under existing principles, an adviser has to accurately describe its methods and services and avoid misleading clients, which leaves each firm, typically in consultation with counsel, to decide for itself whether its AI use is material enough to mention. The report said therefore its 6% adoption figure should be seen as a floor on actual adoption, not a count of firms that crossed some regulatory line.

“Astraeus believes it captures firms at the leading edge and shows how they position AI with clients and regulators. This report provides a baseline for measuring meaningful AI adoption,” the report said, adding that its brochure analysis understates real adoption.

Across the 341 brochures using the term “AI” directly, 42% were dominated by risk language, 55% balanced risk against benefit, and only 4% led with benefits, the report said. Nearly 40% of disclosing firms went out of their way to state that AI does not make investment decisions or substitute for human review. Astraeus read the pattern as evidence that legal counsel, not marketing or the C-suite, is driving what firms say about AI to regulators.

Where AI shows up at all, it’s mostly in the back office. Close to half of disclosing firms cited AI for tasks like meeting notes, CRM updates and document drafting, and about a third cited it for investment research. Fewer than 5% described AI as a direct input into asset allocation or security selection.

AI use also tracked closely with investment complexity. Firms disclosing AI use were about twice as likely to offer private equity as those that didn’t, at 47% versus 24%, and three times as likely to offer private credit. Firms citing AI were also more likely to offer tax-optimized or direct indexing strategies, which the report attributed to AI’s usefulness in pulling structured data out of unstructured documents like private placement memoranda and due diligence questionnaires, rather than any expansion of AI’s role in actual portfolio decisions.