Financial advice has a math problem. More households want advice than ever, trillions of dollars are about to change hands between generations, and the number of advisors is flat to shrinking. AI is the obvious answer, and most advisors are already using it to take notes. But regulators have made clear that the rules on supervision, books and records, marketing claims and client data apply to AI just as they do to everything else. For an independent firm the real question is not whether to use AI. It is whether you can use it at scale and still pass an exam.
The capacity problem AI is supposed to solve
McKinsey projects that about 110,000 advisors, roughly 38% of the workforce, will retire over the next decade. Advised relationships, meanwhile, could grow from 53 million to as many as 71 million by 2034. At today's productivity, that leaves the industry roughly 100,000 advisors short. Cerulli estimates about $124 trillion in U.S. wealth will transfer through 2048, much of it to heirs who may not keep their parents' advisor.
The time is not lost in meetings. Kitces Research found the average advisor spends about 8.8 hours a week meeting current clients, against roughly 5.3 hours of prep, 6.6 of analysis and 6 of follow-through. That is about two hours behind the scenes for every hour in front of a client. The firms that grow without hiring do not cut client time. They take the work around each client hour away from the advisor.
Where advisors actually use AI in 2026
Adoption is broad. AssetMark's 2026 survey of 400 U.S. advisors, conducted in June and July, found 85% have adopted AI-integrated tools to some degree, 80% expect to use them more over the next year, and 55% save at least four hours a week. Meeting notes and summaries are where most firms start. FINRA's 2026 oversight report likewise found summarization and information extraction to be the top generative AI use among member firms.
Firms are early, not mature. Cerulli's 2026 study of RIA firms, done with Vista Equity Partners, placed half of them in the lowest AI maturity tier and only 12% in the top tier. The same firms mostly plan to hire more junior advisors and service associates over the next two years. AI is being used to expand capacity, not replace people.
Trust is the limit. Advisors are adopting AI for internal work far faster than for anything a client sees, and for good reason. FINRA's report flags AI agents that act without human approval or beyond their authorized scope, multi-step reasoning that is hard to trace, and exposure of sensitive data as emerging risks. Regulators are watching client-facing AI and compliance closely.
A meeting summary, a follow-up email and a prospecting message are not just productivity output. To an examiner, they are records.
What regulators expect from AI in an advisory firm
No AI-specific rule, but the existing ones apply. In June 2025 the SEC formally withdrew its 2023 proposal on predictive data analytics, which would have created AI-specific duties for advisers and broker-dealers. That did not create a gap. The SEC's fiscal 2026 examination priorities say examiners remain focused on firms' use of AI technologies and automated tools, and on the controls firms use against AI-related cyber risks. FINRA's 2026 Regulatory Oversight Report says its rules apply to generative AI as they do to any technology, and points to supervision under Rule 3110, communications, recordkeeping and fair dealing.
AI output is a record. The Advisers Act books-and-records rule (Rule 204-2) defines records to include "transcribed information of any type," and practitioners at firms like Cooley read it to reach AI-generated notes, summaries and transcripts. FINRA's report treats AI output under its existing recordkeeping rules. Regulators have already shown how seriously they take records: by early 2024, SEC and CFTC fines for off-channel communications and recordkeeping failures had passed $2.6 billion. An AI tool that produces client communications outside the firm's archive is the same problem in a new form.
Say only what the AI actually does. In March 2024 the SEC settled charges against two advisers, Delphia and Global Predictions, for "AI washing," claiming AI capabilities they did not have, with $400,000 in combined penalties. Advisers should make sure their marketing, Form ADV and client communications describe how the firm actually uses AI.
Client data and your vendors. The amended Regulation S-P requires a written incident response program, notice to affected customers within 30 days of a breach involving sensitive information, and oversight of service providers. Smaller firms had to comply by June 3, 2026. Every AI tool that touches client information is a service provider. A consumer chatbot one associate uses on their phone is a service-provider arrangement the firm has no record of, and consumer tiers of the major chatbots generally train on what is typed in unless the user opts out. Financial services breaches averaged $6.29 million in IBM's 2026 Cost of a Data Breach report.
A practical AI checklist for advisory firms
1. Inventory every AI tool, including the ones nobody approved. List what each one touches, whether it trains on your data, how long it keeps it, and where its output is stored.
2. Capture AI output in your books and records. Notes, summaries, drafts and messages should land in the CRM and archive automatically, not in a vendor's app.
3. Keep a human on every client-facing word. AI drafts the follow-up letter, the meeting summary and the prospecting message. An advisor or compliance reviews and sends it.
4. Describe your AI use accurately. Check your website, ADV and pitch materials against what your tools actually do.
5. Fold AI vendors into your Regulation S-P program. Each AI vendor belongs in your service-provider records and your incident response plan.
How Acclivity Labs helps
Acclivity Labs builds and runs AI agents for independent advisory firms and other small and mid-sized businesses. Our Scribe agent takes the meeting notes, and it does not work alone. It runs on one private platform where every agent shares one CRM, one approval model and one audit trail:
Client service. Meeting prep packs from your CRM and custodian data before the meeting, and our Scribe agent's notes and action items after it, filed straight into your CRM and archive, follow-up letters drafted for the advisor to approve, paperwork prepared and tracked, and signatures collected by eSign. A receptionist agent answers calls and books reviews.
Growth. Prospect and center-of-influence follow-up from a compliance-approved message library, at human pace, with every touch recorded. Websites, search and AI-answer visibility from Acclivity Digital Marketing, our affiliated agency, so prospects who ask an AI for an advisor in your niche can find you. Content is drafted for compliance review before anything posts.
Planning, servicing and the back office. Policy and plan servicing requests, review and renewal reminders, planning scenarios in the client's own numbers, firm books, billing reconciliation and licensing and CE tracking.
Compliance and security. Help building and maintaining the written incident response program, service-provider records and 30-day notification process that Regulation S-P requires, with managed IT from our affiliated partner.
Built for examiners. Our agents run on our own Nvidia DGX hardware using open-weight models. Client data stays in your own tenant, under your retention policy, and is never used to train a model. A hosted frontier model is used only where a task needs it, and only under commercial no-training terms. Every draft, message and approval is logged, so you can show an examiner what the AI produced, who changed it and who sent it.
The advisory firms that grow through the coming advisor shortage will be the ones that give their advisors back the two hours around every client hour, and can prove to an examiner exactly how they did it.