The part of advisory work that eats your calendar is preparation and follow-through: reviewing the file before a meeting, writing up what was discussed, answering the same six questions by email, and producing content your compliance team will approve. AI cuts each of those from an hour to minutes, working from material you already have.
What it cannot do is advise. A chat model does not know the client's full picture, is not a fiduciary, and will confidently state a contribution limit that changed last year or describe a product feature that does not exist. Anything it produces is a draft for you, never a recommendation for a client, and it never goes out without the same review your own writing gets.
SEC and FINRA rules on communications, supervision, and books and records apply whether a person or a model wrote the first draft. Client nonpublic personal information is protected under Regulation S-P and the Gramm-Leach-Bliley Act. Never paste client financial data into a consumer AI tool unless your firm has approved that tool and its data terms.
Quick wins this week
- Paste three de-identified meeting notes for one client type and ask for a one-page prep brief format you can reuse: what changed, open items, questions to ask, documents to bring.
- Take the explanation you give of one concept, for example why you rebalance, and ask for three versions: one for an engineer, one for a retiree who wants the short answer, one for a couple in their thirties. Send them to compliance for pre-approval as templates.
- Upload a fund prospectus or a new plan document to NotebookLM and ask about fees, share classes, or distribution rules with page references. Then read the page.
- Turn a recorded meeting (with consent, through a firm-approved tool) into a summary and action list, then edit it before it becomes the CRM note.
What AI can do for financial advisors, task by task
Meeting prep briefs
Export the client's recent CRM notes, the last review summary, and the agenda items, remove the name and account numbers, and ask for a prep brief: what changed since last time, open items, questions to ask, and topics the client raised before. Check it against the actual file; the model fills gaps with plausible details that are not in your notes.
Post-meeting summaries and follow-up letters
Use a firm-approved recorder with the client's consent, or paste your own notes, and ask for a summary in your firm's format plus a follow-up email: what was discussed, what was decided, what you will do, what the client will do. Review every line for accuracy, then send through an archived channel. The summary is a record; treat it that way.
Explaining concepts without giving advice
Ask for a plain-English explanation of a concept the client asked about, with a worked example using round numbers and no product names. Add your own context, strip anything that reads as a recommendation, and keep figures as placeholders you fill from current sources. The tool explains how something works; you say what it means for this client.
Client education content and marketing
Draft newsletters, webinar outlines, and social posts from your own point of view, then run each through compliance pre-approval as you would any retail communication. Tell the model to avoid performance claims, guarantees, testimonials, and predictions. The SEC Marketing Rule and FINRA Rule 2210 apply to the finished piece regardless of who drafted it.
Reading prospectuses, plan documents, and estate documents
Upload the document to NotebookLM or Claude and ask specific questions with page references: expense ratios by share class, surrender schedules, required distributions, trustee powers. Then read the cited page yourself. For a client's own estate documents, use only a firm-approved tool and confirm your engagement permits it.
Planning scenario framing
Before you build a plan in eMoney, MoneyGuide, or RightCapital, describe the household without identifiers and ask what data to gather, which assumptions matter most, and what scenarios to model. The planning software runs the numbers; the model helps you structure the conversation. Never ask it to project returns or run the plan itself.
Prompts for financial advisors
Replace the bracketed placeholders, paste into any chat assistant, and iterate on the result.
Meeting prep brief from CRM notes
You are an associate advisor preparing a review-meeting brief for a lead advisor. Below are de-identified CRM notes and the last review summary for a client household ([HOUSEHOLD DESCRIPTION, E.G. COUPLE IN THEIR FIFTIES, TWO KIDS IN COLLEGE]). Produce a one-page brief: 1) what changed since the last meeting, 2) open items with the date each was first raised, 3) questions to ask, 4) topics the client raised more than once, 5) documents to bring. Use only facts in the notes; if something is unclear, list it under Questions rather than assuming. No recommendations. Notes: [PASTE DE-IDENTIFIED NOTES]
Tip: Anything in the brief that is not in your notes is a hallucination; that check takes thirty seconds.
Post-meeting summary and follow-up email
You are an advisor's assistant. From the meeting notes below, write 1) a CRM summary with these headings: Discussed, Decided, Advisor Actions, Client Actions, Next Meeting, and 2) a follow-up email to the client under 200 words in a warm, direct tone that restates the actions and dates. Do not add anything that is not in the notes. Do not describe any product, projected return, or tax outcome. Leave [DATE] placeholders where dates were not stated. Meeting notes: [PASTE YOUR NOTES OR THE APPROVED TRANSCRIPT SUMMARY]
Tip: Send the email through your archived email system, never from the AI tool.
Explain a concept three ways
You are a financial educator writing for an advisor's clients. Explain [CONCEPT, E.G. SEQUENCE-OF-RETURNS RISK OR HOW A ROTH CONVERSION WORKS MECHANICALLY] in three versions of under 150 words each: 1) for an analytical client who wants the mechanism, 2) for a client who wants the short answer, 3) for a young couple new to investing. Use round hypothetical numbers, no product names, no predictions, no recommendations, and no statements about tax rates or limits; write [CURRENT FIGURE] wherever a number would go. End each version with: Ask your advisor how this applies to you.
Tip: Get the three versions pre-approved once as templates; then they are reusable all year.
Compliance-ready newsletter draft
You are a content writer for a registered investment adviser. Draft a client newsletter article of about 400 words on [TOPIC] in the first person from the advisor's perspective. Rules: no performance claims, no guarantees, no predictions about markets or rates, no testimonials, no specific product or security names, no tax or legal advice. Use hypothetical examples labeled as hypothetical. Use sentence-case headings and plain American English. End with a placeholder line for the firm's required disclosure: [FIRM DISCLOSURE].
Tip: Your compliance reviewer still reads every word; the rules in the prompt just reduce their markups.
Questions to gather before building a plan
You are a paraplanner supporting a financial planner. A prospective client household is described below without identifiers. List 1) the data points needed to build a comprehensive plan, grouped by category, 2) the five assumptions that will most affect the outcome and why, 3) the scenarios worth modeling, 4) questions to ask that clients often do not volunteer.
Do not project returns, suggest allocations, or recommend products.
Household: [AGES, EMPLOYMENT, ROUGH ASSET AND INCOME PICTURE, STATED GOALS, KNOWN CONCERNS]Tip: Take the list into your planning software; the model never runs the plan.
Want a prompt for something else? Use the Prompt Builder.
Skills to build
Drafting without advising
Why: The line between explaining a concept and recommending an action is the line between a helpful tool and a compliance problem, and you need to see it in every draft.
How: Write no recommendations, no predictions, no product names into every saved prompt. Read each output asking: could a client act on this sentence alone?
Keeping the record straight
Why: Meeting summaries, follow-up emails, and content are books and records. If AI drafted them, they still need to live in your archived systems.
How: Draft in the AI tool, finalize in your CRM and email. Ask compliance how they want AI-assisted drafts documented and follow that exactly.
Checking figures against current sources
Why: Contribution limits, tax brackets, required distribution ages, and Social Security rules change, and the model's training data is behind.
How: Prompt for [CURRENT FIGURE] placeholders. Keep a one-page current-year reference from your custodian or planning software and fill from that.
Working from uploaded documents
Why: Prospectus and plan-document questions are answered best by a tool that reads the actual document and shows the page, not by one recalling similar documents.
How: Use NotebookLM or a Claude Project per document type. Always open the cited page before you rely on it.
Separating what can be automated from what cannot
Why: Recurring outreach and review scheduling can run on rails; the judgment calls cannot. Knowing which is which is the whole productivity gain.
How: List your monthly touchpoints, mark the ones that never require judgment, and template or automate those first through firm-approved systems.
Tools worth knowing
Claude
A careful writing and analysis assistant that shines on long documents.
ChatGPT
The general-purpose AI assistant most of your coworkers already use.
Microsoft Copilot
AI inside Word, Excel, Outlook, and Teams, with your company's data protections.
NotebookLM
A research notebook that only answers from the sources you give it, with citations.
Zoom AI Companion
The meeting notetaker and assistant already built into your paid Zoom account
Canva
Design for non-designers, with Magic Studio AI for text, images, video, and resizing.
Cautions for financial advisors
A chat model is not a fiduciary, does not know the client's full situation, and is not subject to Regulation Best Interest or the Advisers Act. Never forward its output as a recommendation, and never let a client-facing chatbot answer questions about what to buy, sell, or hold. Explanations of how something works are fine; what this client should do comes from you.
FINRA Rule 2210 and the SEC Marketing Rule govern the finished communication regardless of who drafted it, and FINRA has reminded firms that its existing rules apply to generative AI. Retail communications still need principal review where required. Follow-up emails and summaries are records under Exchange Act Rule 17a-4 or Advisers Act Rule 204-2; keep them in archived channels, and ask compliance whether AI prompts and outputs about client matters must be retained too.
Nonpublic personal information (names tied to holdings, account numbers, income, SSNs) is protected under Regulation S-P and GLBA. Consumer AI tiers may retain or train on inputs. Use only tools your firm has approved, with terms that exclude your data from training, and de-identify anyway. Never paste client financial data into a consumer AI tool unless your organization has approved it.
AI note-takers record. Several states require all-party consent, and your firm's policy may require specific disclosure language and archiving of transcripts. Use the recorder your firm approved, get consent on the record, and review the summary before it becomes the CRM note; recorders mis-hear numbers and names.
The model will state last year's contribution limit, misdescribe a fund's share classes, or invent a feature of an annuity. Treat every figure as a placeholder and every product description as unverified until you have read the source document.
Your 30-day plan
- Week 1: Ask compliance which AI tools are approved, in which tiers, and how AI-assisted drafts should be documented. If there is no answer yet, use approved tools only on non-client content.
- Week 1: Do three quick wins on non-client material: concept explanations, a newsletter outline, and a prep-brief format.
- Week 2: Run your next five meeting preps through the brief prompt with de-identified notes. Score each for anything invented.
- Week 2: Use a firm-approved recorder, with consent, on three meetings and compare the AI summary to your own notes. Log every mis-heard figure.
- Week 3: Build a Claude Project or custom GPT with your tone, disclosure placeholders, and the no-recommendations rules. Get your templates pre-approved.
- Week 4: Pick one recurring touchpoint to template or automate through firm-approved systems, then show your team the results and the errors you caught.
Frequently asked questions
Can financial advisors use ChatGPT with clients?
Is AI-generated content allowed under FINRA and SEC rules?
Will AI replace financial advisors?
Can I use an AI note-taker in client meetings?
Should I tell clients I use AI?
Terms used on this page
Related roles
- AI for AccountantsAI drafts the flux commentary, writes the XLOOKUP, explains the K-1 to your client in plain English, and reads the 40-page lease for you. You still own the numbers, the judgment, and the signature.
- AI for Insurance AgentsAI explains replacement cost versus actual cash value to a client in two paragraphs, drafts the rate-increase email you dread, and builds a quote comparison from three dec pages. The advice, the license, and the E&O exposure stay yours.
- AI for Financial AnalystsAI writes the first pass of your variance commentary, audits your model for hard-codes and sign errors, and turns a 10-K into a table with page references. You decide what the numbers mean and what to tell the CFO.
- AI for Tax PreparersAI drafts the organizer letter, the CP2000 explanation, and the note about why the refund shrank, and it reads a 900-page bill so you can ask it questions. Every threshold, date, and citation still gets checked against this year's source before it touches a return.