Skip to main content
Menu
Wealth Management Technology

What is an AI CRM for wealth management?

Diagram of an AI CRM loop around a household record: capture, understand, suggest and act cards, with permissions and audit log chips, on a soft pastel gradient.

An AI CRM can read, update and act from the client record within each user's permissions. Here is what it should do, what it needs underneath and how to govern AI notes.

An AI CRM for wealth management is a client relationship system where AI can read the household record, add to it and act from it on a team member's behalf. In practice, that means logging meetings and emails, summarizing what changed for a client, suggesting next steps and creating tasks, all within the permissions of the person using it. How useful it is depends less on the model than on what the CRM holds. If custodian, planning and communication data sit in other systems, the AI works from a partial picture of the client.

That makes choosing an AI CRM a data and controls decision as much as a feature decision. This guide covers what an AI CRM should do, what it needs underneath, how recordkeeping applies to AI-generated notes, and the questions to ask before you buy one.

Why the CRM is where AI lands first

For many advisory firms, the CRM is already the center of the tech stack. In Kitces Research's 2025 AdvisorTech study, 48% of advisors said their CRM is the core system their other tools revolve around, ahead of financial planning software (17%) and portfolio management platforms (14%). The same study found the median advisory firm uses 12 technology applications to support 20 business functions (Kitces Research).

AI has followed the CRM. Kitces reported in April 2025 that AI meeting note tools went from essentially nonexistent to the fastest-growing category on its AdvisorTech Map in about two years, with about 18% of advisory teams adopting one. Advisor-specific tools log meeting notes in the client's CRM file, often prompt follow-up tasks, and draft a recap email for the advisor to edit (Kitces). Kitces research also found that these industry-specific tools led generic notetakers on advisor satisfaction, largely because of the CRM-integrated work that follows a meeting (Kitces).

A meeting is one input, though. A notetaker connected to the CRM still sees only what was said on the call and what the CRM already knows. Transfers, holdings changes and planning updates arrive elsewhere.

Four jobs an AI CRM should do

1. Capture

Record client interactions without retyping: approved email and calendar activity, meeting notes, calls and team conversations, attached to the right household with a source and a timestamp.

2. Understand

Summarize what changed for a household since the last touch, such as new account activity, open tasks, outstanding documents and life events mentioned in conversation. Summaries should show where information came from so the advisor can check it.

3. Suggest

Flag what needs attention and propose next steps, such as a follow-up after a meeting or a review when a client's circumstances change. Each suggestion should be easy to accept, edit or dismiss.

4. Act

Carry out supported tasks when a person asks, such as creating a task, logging an activity note or updating a field. This is where time savings grow and where controls matter most, because the AI is now changing the record.

These jobs match how advisors say they want to use AI. In the Kitces 2025 study, 57% of advisors expected AI to expedite key tasks and 15% expected it to help them initiate tasks, while 28% wanted it to fully automate tasks. The report concludes that AI solutions should build in human-in-the-loop capabilities rather than lead with hands-off automation.

What an AI CRM needs underneath

One household record with custodian and planning data

If accounts and transfers live in a custodian portal and goals live in planning software, an AI asked what changed for a household can only answer from the CRM's slice. Ask which sources feed the record, which direction data flows, and how stale data is flagged. Our guide to reducing manual CRM data entry covers assigning an authoritative source to each field.

A shared timeline the whole team uses

Advisors, associates, operations and service staff all touch the same household. When their conversations live in personal inboxes and side chats, the AI misses them, and so does the next person who picks up the client. A shared timeline gives people and AI the same context.

Permissions the AI inherits

An AI that can read every record in the firm can surface information its user should never see. Limit the AI to what the person using it can access, with firm-level settings to narrow it further.

An audit log that covers AI actions

When AI creates tasks or changes fields, supervisors need to see what it did, on whose instruction and when. Human and AI actions belong in the same reviewable log. FINRA's 2026 oversight report lists risks specific to AI agents, including acting without human validation, acting beyond the user's intended scope and authority, and multi-step reasoning that is hard to trace. It suggests firms consider how to monitor agent access and data handling, where to keep a human in the loop, and how to track agent actions (FINRA).

Treat AI-generated notes as records

For registered investment advisers, the books and records rule casts a wide net. Cooley's April 2026 analysis points out that Rule 204-2 defines a record to include "transcribed information of any type," which sweeps in AI-generated notes, summaries and transcripts (Cooley; 17 CFR 275.204-2). Cooley recommends making policy decisions in advance, by category of call. For some call types, such as calls with counsel, AI notetakers may not be appropriate. For others, firms can set clear rules about what is retained and in what form. Cooley also warns that retaining everything is not automatically the conservative approach, and that deleting a transcript that was a required record may itself be a violation.

For broker-dealers, FINRA says its rules are technology neutral and apply to generative AI as they do to any other tool, including features embedded in existing products (FINRA Regulatory Notice 24-09). The 2026 oversight report adds that generative AI use can implicate supervision, communications, recordkeeping and fair dealing obligations.

Before turning on AI capture in a CRM, decide:

  • Which meeting types AI may record and summarize, and which it may not.
  • Whether the transcript, the summary or both become part of the client record.
  • Who reviews AI summaries before they are final.
  • Where the vendor stores AI output, for how long, and how the firm can produce it on request.
  • Whether client data is used to train models.

Confirm the requirements for your firm with compliance counsel.

Questions to ask an AI CRM vendor

  1. Which data sources feed the household record, and which are live for your custodians and market?
  2. Can the AI answer questions using custodian and planning data, or only notes and contacts?
  3. Which actions can the AI take, and does it confirm what it did?
  4. Can the AI see or do anything its user cannot?
  5. Are AI actions written to the same audit log as human actions?
  6. Where is data processed, which model provider is used, and is client data used for training?
  7. If you keep your existing CRM, does data sync in both directions?

How OneVest Core works as an AI CRM

OneVest Core is the foundation package of OneVest's AI-native wealth platform, built for firms serving 100 or more clients. According to OneVest's product pages, Core includes:

  • A CRM built for wealth, with household, client and contact management, opportunities and pipeline tracking, and email and calendar sync.
  • Custodian, CRM and planning data in every client record. Current integrations include Schwab, Fidelity and Apex for custodial data, Orion and Envestnet for portfolio data, and Nylas for email and calendar. Firms that keep Salesforce can use a bi-directional sync.
  • Pulse, a shared feed of client activity and team conversations, with AI that summarizes activity and surfaces what needs attention, plus audit logging. Slack and Microsoft Teams integrations bring team chat into the same view.
  • AI that monitors client changes and suggests next steps, including customer life events detection, and the OV Bot AI assistant.
  • APIs, developer tools, MCP and webhooks, so firms can connect compatible AI tools with scoped access.

According to OneVest's AI FAQ, AI in the platform carries out supported actions from an instruction in chat, such as creating tasks and logging activity notes, and then confirms what it did. It can only do what the person using it is allowed to do, organization settings can narrow that further, and every action by a person or AI is recorded in the audit log. OneVest does not train models on client data, and data is processed in the US or Canada based on the firm's location.

Core covers the internal relationship record. The branded client portal, client messaging and sentiment scoring are part of Engage, and account opening, money movement and case management are part of Operate. The Compare packages page shows what each package includes.

Start with one workflow

Pick one advisor team and one recurring job, such as preparing for annual reviews. Measure how long it takes today and how often notes are missing or late. Turn on AI capture and summaries for that team, review the output against your records policy for a few weeks, and then add suggested next steps and supported actions. For more places AI fits in a wealth firm, see our guide on how to use AI in wealth management.

Explore OneVest Core

Sources