73% of employee advisers actively use AI (up from 44% a year earlier) and independent adviser adoption jumped from 19% to 42%, per J.D. Power's 2026 Financial Advisor Satisfaction Study.Source ↗
Only 9% of advisors say they don't use generative AI at all, and 82% report their firm now has a formal AI policy — up from 47% a year earlier, per Advisor360°'s advisor survey.Source ↗
FINRA's 2026 Regulatory Oversight Report found summarization and information extraction is the most common generative AI use case at member firms — and reminds firms that supervision, communications, and recordkeeping rules all apply to AI output.Source ↗
planningClaudeChatGPT

Client review meeting prep briefs in ten minutes

Annual review season means reconstructing each client's situation from scattered CRM notes, last meeting's action items, and whatever changed in their life since. Surveys show meeting prep and admin are among the first things advisers hand to AI — 59% automate scheduling, calendar, and prep work — because a structured brief beats skimming two years of notes in the car.

Prompt
You are an experienced paraplanner preparing an advisor for a client review meeting. Using ONLY the information I provide below, build a one-page prep brief.

Client situation (anonymized — no names or account numbers): {{client_summary}}

Notes and action items from the last meeting: {{last_meeting_notes}}

Primary goal for this meeting: {{meeting_goal}}

Produce:
1. A suggested 45-minute agenda with time blocks.
2. Open action items from last time — theirs and ours — with status unknowns marked [CONFIRM].
3. Life events or dates in the notes that warrant a follow-up question (job change, new grandchild, business sale, RMD age).
4. Five questions the advisor should ask, ordered by importance to the stated meeting goal.
5. Topics the client is likely to raise, based only on concerns recorded in the notes.

Constraints: do not invent account values, holdings, or facts not present in what I gave you — mark gaps as [MISSING] instead. Do not recommend any product, allocation, or transaction; this brief is internal preparation, not advice. Plain language, bullets, one page.

Fill in your details and the prompt updates live — then copy.

communicationClaudeChatGPT

Calm client emails when markets drop

A bad week in the markets fills your inbox with variations of "should we be doing something?" — and every reply is a regulated communication. Kitces' analysis of ChatGPT for advisors landed on exactly this workflow: prompting a draft and editing it for accuracy and personalization is faster than composing from scratch, and 53% of advisers now use AI to draft routine client emails.

Prompt
You are drafting an email for a fiduciary financial advisor to send to a client who is anxious about {{market_event}}. Client context, anonymized: {{client_context}}.

Write a reply that:
- Opens by acknowledging the concern specifically — no "I understand your concern" boilerplate.
- Restates, in one short paragraph, why their plan was built to withstand periods like this, using only the context I gave you.
- Explains one relevant concept ({{concept_to_explain}}) in plain English, no jargon, max three sentences.
- Closes with a concrete offer: a call this week to walk through their plan.

Hard rules: under 250 words. No predictions about where markets are headed. No guarantees, no "markets always recover," no promissory language of any kind — this must survive review under FINRA Rule 2210's prohibition on exaggerated or promissory statements. Do not recommend buying, selling, or changing anything. Do not cite statistics I did not provide. Where a plan-specific number belongs, insert [ADVISOR: fill in] rather than inventing one. Warm, steady, professional tone — like a doctor with good bedside manner, not a salesperson.

Fill in your details and the prompt updates live — then copy.

automationClaudeChatGPTCopilot

Turning meeting notes into CRM records, tasks, and follow-ups

The 30 minutes of admin after every client meeting — CRM note, follow-up email, task list — is where AI adoption in wealth management actually started. Kitces research tracks AI notetakers as one of the highest-satisfaction tool categories for advisors, and firms like the $6B+ RFG rolled out purpose-built tools (Zocks, Jump) firm-wide. Even with plain ChatGPT or Claude, raw notes become structured output in one pass.

Prompt
You are a financial advisor's associate processing raw notes from a client meeting. The notes are anonymized. Convert them into three deliverables.

Raw meeting notes: {{raw_notes}}

Firm task categories to use: {{task_categories}}

Deliverable 1 — CRM meeting record: date/attendees placeholder, topics discussed, decisions made, risks or concerns the client raised, and any disclosures given. Use the exact wording from my notes for anything the client said about risk tolerance or objectives — do not paraphrase those.

Deliverable 2 — Client follow-up email draft: warm, under 150 words, recapping decisions and next steps only. No new recommendations.

Deliverable 3 — Task list: each task with an owner (advisor / staff / client), a suggested due date relative to today, and one of my firm's task categories.

Constraints: include only what is actually in the notes; mark ambiguous items [CONFIRM] rather than guessing. If anything in the notes sounds like a NEW recommendation not previously documented, flag it separately as NEEDS ADVISOR REVIEW — do not put it in the client email.

Fill in your details and the prompt updates live — then copy.

writingClaudeChatGPTGemini

Newsletter and LinkedIn drafts that survive compliance review

Marketing is one of the top three things advisors use generative AI for, and it's also where the regulatory exposure is sharpest: the SEC's first AI-related Marketing Rule actions fined two advisers $225,000 and $175,000 for overstated AI claims, and FINRA Rule 2210 requires principal approval of retail communications before first use. The workable pattern is drafting with compliance constraints baked into the prompt — then routing everything through review anyway.

Prompt
You are a financial services marketing writer who knows the SEC Marketing Rule and FINRA Rule 2210 cold. Draft {{content_type}} on the topic of {{topic}} for this audience: {{audience}}.

Style: educational, plain English, specific and useful — the reader should learn something they can act on in a conversation with their advisor. No hype. Banned framing: fear-based urgency, "secrets the wealthy know," anything that reads like a pitch.

Hard compliance constraints:
- No performance claims, projections, or forward-looking return statements.
- No testimonials, endorsements, or hypothetical performance.
- Never use "guarantee," "ensure," "will grow," "outperform," or similar promissory language.
- No specific securities, funds, or products by name.
- Do not invent statistics, dates, or dollar limits. Where a figure is needed (contribution limits, tax brackets), write [SOURCE NEEDED: description] so I can insert the verified current number.

After the draft, append a section titled COMPLIANCE FLAGS listing every sentence a reviewing principal might question and why — even borderline ones. Then suggest one compliant disclosure line appropriate to the content.

Fill in your details and the prompt updates live — then copy.

analysisClaudeChatGPT

Extracting the fine print from annuities, plan documents, and prospectuses

A client brings you an annuity contract, a 90-page employer benefits guide, or a fund prospectus and asks "is this any good?" FINRA's 2026 Regulatory Oversight Report found summarization and information extraction is the single most common generative AI use case at member firms — because pulling fees, surrender schedules, and exclusions out of dense documents is exactly what these models do well, as a first pass you then verify.

Prompt
You are an analyst supporting a fiduciary financial advisor. I will paste text from {{document_type}}. The document has been anonymized. Extract and organize the following, focused on {{extraction_focus}}:

1. A table of every fee, charge, and expense mentioned — name, amount or formula, when it applies, and the EXACT quoted language from the document next to each entry.
2. Key dates and schedules (surrender periods, vesting, rate reset dates, deadlines).
3. Riders, features, or options and what each costs.
4. Exclusions, limitations, and conditions buried in the definitions or fine print.
5. A plain-English summary at an 8th-grade reading level: what this product does, what it costs, and what the owner gives up.
6. A "CONFIRM WITH ISSUER" list: every ambiguity or term where the text is unclear or seems to conflict with itself.

Hard rules: extract only what is in the pasted text. If something I asked for is not present, write NOT FOUND — never infer or estimate a missing number. Quote source language for every figure. This is a first-pass extraction; I will verify each item against the original document.

Fill in your details and the prompt updates live — then copy.

creativeChatGPTClaudeGemini

Building client education seminars and webinar content

Educational seminars and webinars remain a workhorse for client retention and referrals, but building a 45-minute session from scratch — outline, slides, handout, promotion — is a week of nights. CFP Board's Generative AI Ethics Guide lists content creation and brand-building ideas among the legitimate uses for CFP professionals, provided the output is verified and the professional judgment stays yours.

Prompt
You are a practice-development strategist for an independent financial advisor whose niche is {{niche}}. Design a 45-minute educational seminar on {{topic}}.

Deliver:
1. Three title options — specific and curiosity-driven, no hype words.
2. A segment-by-segment outline with timing: hook (5 min), three teaching segments (10 min each), Q&A framing (10 min). For each segment: the one idea attendees should remember, a story or analogy suggestion, and a discussion question.
3. A one-page handout outline attendees take home — worksheet-style, with blanks they fill in during the session.
4. A 120-word promotional email inviting clients to bring a friend.

Constraints: education only — no product mentions, no return projections, no "strategies that guarantee" anything, nothing that pressures attendees toward a sale. Where current-year figures belong (contribution limits, tax thresholds, Medicare premiums), write [ADD CURRENT FIGURE] instead of a number. Reading level: bright high-schooler. The goal is that attendees leave with one decision they know how to think about, not a pitch they sat through.

Fill in your details and the prompt updates live — then copy.

Common questions from financial advisors

Is it even allowed for a financial advisor to use ChatGPT or Claude?

There is no rule banning it. FINRA's Notice 24-09 takes a technology-neutral stance — the same supervision, communication, and recordkeeping rules apply to AI output as to anything you type yourself. The practical gate is your firm's policy, and 82% of firms now have a formal one, so read it before you paste anything.

Can I put client information into AI tools?

Not into consumer tools. Client names, account numbers, holdings, and financial details are nonpublic personal information under Regulation S-P, and CFP Board's ethics guidance makes confidentiality the first checklist item. Anonymize thoroughly, or use an enterprise deployment your firm has vetted and approved — those exist precisely because the free consumer versions may train on or retain what you enter.

Do AI-drafted emails and marketing need compliance review and archiving?

Yes, exactly as if you wrote them. Client correspondence follows your firm's supervision procedures, retail communications need principal approval before first use under FINRA Rule 2210, and books-and-records retention applies to AI-generated content used in your business. Kitces' compliance analysis puts it simply — once AI output informs a client communication or a decision, it likely requires retention.

Can I rely on AI for actual investment recommendations?

No — and neither can your clients rely on its output as advice. AI output becomes advice only after you have verified it against the client's actual situation and applied your own suitability, best-interest, or fiduciary analysis; that judgment is not delegable. Notably, advisors themselves seem to agree — building personalized financial plans is their least common AI use at just 29%.

Related professions