← All posts
February 14, 2026·11 min read

AI notes for financial advisors in the US: FINRA, Reg BI, and the meeting that wasn't on Zoom

US financial advisors operate under FINRA Rule 4511, Reg BI suitability documentation, and a 2026 regulatory environment where AI use is under active supervision. Cloud AI notetakers create recordkeeping, client confidentiality, and NDA exposure that in-person meetings make worse. On-device documentation changes the compliance equation.

Key takeaways
  • FINRA Rule 4511 requires that books and records be kept for specified periods. Suitability rationale under Reg BI must be documented. AI-generated meeting notes support both when reviewed and retained by the advisor.
  • FINRA's 2026 Annual Regulatory Oversight Report includes the first dedicated AI section — supervision and recordkeeping requirements explicitly apply to AI tools used in client-facing workflows.
  • Cloud AI notetakers processing client meeting audio — which may include NDA-sensitive investment strategy, AUM disclosures, and personal financial data — create data exposure that formal client NDAs may prohibit.
  • In-person client meetings at HNW offices and family offices require a phone-based recording tool. Bot-based notetakers cannot join a meeting in a Manhattan conference room.

Every client meeting a financial advisor holds generates a compliance obligation. Suitability documentation. Recommendation rationale. Client instructions given and acknowledged. Risk disclosures discussed. Under FINRA Rule 4511 and SEC recordkeeping requirements, the written record of what happened in that meeting is not optional — it is a regulatory artifact that must be retained for up to six years and be producible on exam.

The problem is that the meeting itself is what advisors are paid to do. The documentation is what happens after — in the gaps between meetings, at the end of the day, sometimes on the weekend. For a full-service advisor running 8–12 client meetings per week, the documentation burden is a second job grafted onto the first one.

AI documentation changes the ratio. But FINRA's 2026 Regulatory Oversight Report introduced a dedicated section on generative AI — the first time the regulator has formally supervised AI tools in member firms — and the compliance implications for advisors using AI are specific and consequential.

What the documentation requirement actually demands

Post-meeting documentation for a financial advisory client meeting must capture, at minimum: the client's stated objectives and any changes to them, the recommendations made and the suitability basis for each, any client instructions given or declined, the risk disclosures made, and any follow-up commitments from either party.

For investment advisors under Reg BI, the "best interest" standard requires that the rationale for each recommendation be documented — not just the recommendation itself. "Recommended VTSAX based on client's long-horizon growth objective and tolerance for equity volatility" is the defensible note. "Discussed funds" is not.

For broker-dealers, FINRA expects that suitability determinations are documented contemporaneously with the transaction or recommendation. Notes written a week after the meeting from memory, in the event of a client complaint or FINRA examination, carry significantly less weight than contemporaneous records.

FINRA 2026 and the AI governance expectation

FINRA's 2026 Regulatory Oversight Report is the first time the regulator has dedicated a section specifically to generative AI governance. The key expectations for member firms using AI tools:

Supervision, recordkeeping, and communication standards all apply to AI-generated content. An AI-generated meeting summary that is used as the compliance record of a client meeting is subject to the same accuracy and retention requirements as a human-written note.

The 2026 Report specifically addresses "prompt and output logging" and "version tracking" for AI tools. For firms using cloud-based AI notetakers, this creates a question: are the AI outputs retained in a system that satisfies FINRA's recordkeeping requirements, or does the notetaker generate a note that the advisor then copies into their CRM and the original AI output is lost?

On-device AI processing, where the note is generated on the advisor's device and immediately entered into the compliant record-keeping system, avoids the intermediate step. There is no cloud-hosted AI output to track — the note goes from device to system without a third-party server in between.

The client meeting that can't be recorded to the cloud

Many high-net-worth client meetings involve information that clients expect to be handled with discretion that goes beyond contractual obligation. Estate planning discussions. Family dynamics around wealth transfer. Business sale conversations. Performance conversations during market stress.

A meeting recording sent to a cloud transcription service — even one with a signed BAA or DPA equivalent — is a meeting recording on someone else's server. The advisor's NDA and the firm's privacy policy cover the note in the CRM. They may not speak to the audio file and the AI-generated transcript sitting in a cloud vendor's infrastructure.

For advisors who serve clients with specific privacy expectations — family offices, HNW individuals, executives with public company affiliations — the question of where the audio goes is a relationship management question as much as a compliance question. Many advisors report that clients are more comfortable knowing that meeting recordings are processed on-device and not transmitted to a third-party server.

Kuulo records the client meeting on the advisor's iPhone, processes the audio entirely on-device, and generates the structured meeting summary without any audio reaching an external server.

The in-person meeting that video-only tools miss

The majority of AI meeting notetakers are built for video calls: Zoom, Teams, Meet. They join as bots, transcribe the video call, and generate summaries.

Wealth management client relationships are maintained in person. The quarterly review meeting. The estate planning session with the client's attorney. The onboarding meeting at the client's home or office. The informal coffee that turns into a material portfolio discussion.

These meetings happen in rooms, not on screens. A Fathom or Otter bot cannot join them. Kuulo records from the iPhone microphone in any environment — the office, the client's home, a restaurant, a conference room at the custodian's headquarters. Wherever the meeting happens, the documentation workflow runs.

The structured advisory note

Kuulo's Financial Advisory Meeting template generates notes structured around: client objectives reviewed, portfolio performance discussed, recommendations made and rationale, client instructions given, risk disclosures acknowledged, and action items with owners and deadlines.

This structure maps directly onto what suitability documentation requires. The advisor reviews and edits for accuracy, adds any nuances not captured in the recording, and signs off. The resulting note goes into the CRM as the meeting record. Total post-meeting time: 8–12 minutes rather than 30–45.

For an advisor with 10 client meetings per week, that is 3–5 hours recovered weekly — time that returns to prospecting, to relationship development, or simply to being present for the next meeting rather than still documenting the last one.

Supervisory controls and the compliant workflow

For advisors at registered broker-dealers, supervisory procedures require that AI-assisted notes be reviewed by a supervisory principal before entry into the compliance record. This workflow is exactly what Kuulo supports: the AI generates a draft note, the advisor reviews and attests to its accuracy, and the approved note becomes the record. The AI draft is not the record — the advisor's reviewed and attested version is.

This is the same distinction FINRA draws between an AI-generated output and a supervised, attested document. Kuulo does not file notes. It generates drafts that the advisor reviews, edits, and enters into the compliant record-keeping system under their own attestation.

Frequently asked questions

Do AI meeting notes comply with FINRA recordkeeping rules?

AI-generated meeting summaries that are reviewed, confirmed accurate, and retained by the advisor can satisfy FINRA Rule 4511 recordkeeping requirements for client interaction records. The key is that the advisor attests to the accuracy of the record and it is retained in the firm's books and records system.

What does FINRA's 2026 AI guidance say about meeting notes?

FINRA's 2026 Annual Regulatory Oversight Report includes the first dedicated AI section, noting that supervision and recordkeeping requirements apply to AI-generated outputs used in client-facing workflows. Advisors using AI for suitability documentation should ensure the outputs are reviewed, supervised, and retained.

Can financial advisors use AI notetakers for in-person client meetings?

Yes — if the tool doesn't require a video call platform. Bot-based AI notetakers (Fireflies, Otter, Fathom) cannot join in-person meetings. Kuulo records from the iPhone in any setting — the HNW client's office, the advisor's conference room, a restaurant lunch — generating structured suitability documentation from the conversation.

Is client meeting audio subject to financial advisory NDAs?

Many HNW client relationships involve NDAs that cover the content of financial discussions, investment strategies, and portfolio information. Transmitting client meeting audio to a cloud AI server may violate these NDAs regardless of the vendor's privacy policy. On-device processing keeps audio within the advisor's custody — consistent with the NDA's data perimeter.

Try Kuulo

On-device AI notes, private by design. Free for iPhone and Mac.

Get the app