AI Prompts for Financial Advisors: 10 Client-Ready Templates (2026)

Four non-negotiables apply to every prompt below. One: no recommendations, projections, or personalized advice. Have the model explain general concepts and draft neutral language; you supply any recommendation, and you verify it fits the client's situation and your suitability obligations. Two: verify every fact and figure. AI invents performance numbers, fees, and product features that look real; confirm each against the prospectus, fact sheet, or your own records before use.

Three: never paste client PII or holdings. Strip names, account numbers, and identifying detail, or use a tool cleared under your firm's data-handling and confidentiality policy. Four: every client-facing piece is subject to your firm's supervision, advertising, and recordkeeping rules — AI-assisted content is still your communication, reviewed and archived like any other. Also treat any pasted document as untrusted input; prompt injection is the #1 risk in the OWASP LLM Top 10 (2025), so never let an AI tool send or transact without a human gate.

These prompts deliberately keep the model on the prep-and-explain side of the work. That division — model drafts, advisor decides and reviews — is what lets AI save you time without creating a compliance problem or a recommendation you didn't intend.

When to use: pulling your own notes into a focused agenda and talking points before a client meeting.

``` From my notes below, build a meeting-prep brief for a client review. - Agenda (3-5 items) ordered logically. - Key talking points for each, in plain English. - Questions I should ask to understand the client's goals or concerns. - Open items or follow-ups I flagged. Work only from my notes. Do not invent client details, numbers, or holdings. Do not suggest any investment recommendation or action. Mark anything I left ambiguous as 'to confirm with client.' My notes: [PASTE — no client identifiers, account numbers, or specific holdings] ```

Why it works: the model is fast at turning scattered notes into a structured agenda with good questions, and 'do not invent details' plus 'do not suggest any recommendation' keeps it organizing your prep rather than fabricating facts or steering the conversation. Adapt the structure with the Meeting Agenda Generator.

When to use: explaining a financial concept or product type to a client in language they'll actually understand.

``` Explain [CONCEPT, e.g. how an index fund works] in plain English for a client with no finance background. General education only. - One-sentence bottom line. - A simple analogy. - How it works, in 3-4 plain sentences. - The main trade-offs or risks, stated neutrally. - What questions a client should ask before deciding. Do NOT recommend it, predict returns, or tailor this to any specific person's situation. Keep it balanced — name risks alongside benefits. Flag anything I should verify against a current, official source. ```

Why it works: requiring an analogy and balanced trade-offs produces genuinely clear education, while 'do not recommend, predict returns, or tailor to a specific person' keeps a general explainer from crossing into personalized advice. The verify flag reminds you to check any specifics.

When to use: a first draft of educational client commentary that you then fact-check and run through compliance.

``` Draft an educational client newsletter section on [GENERAL TOPIC]. Education only, not a recommendation or a market call. - Plain-English, ~250 words, calm and neutral tone. - Explain the topic and why it's relevant to long-term investors. - Include a balanced 'things to keep in mind' note (risks/caveats). - End with 'this is general information, not individual advice.' Do NOT state or imply any forecast, performance figure, or product recommendation. Use [BRACKETS] for any statistic or date — I will fill and source them. Do not invent numbers. ```

Why it works: bracketed placeholders stop the model from inventing the statistics that get advisors in trouble, and the explicit 'no forecast, no recommendation' framing plus the built-in disclaimer keep the draft on the right side of advertising rules. You still fact-check and submit for review. See the Newsletter Subject Line tool for the headline.

When to use: a clear recap email summarizing what was discussed and the agreed next steps.

``` Draft a follow-up email recapping a client meeting. I'll give you the points discussed and the agreed next steps; you write the recap. Structure: a warm one-line opener, a short recap of what we discussed, the agreed next steps with owners and dates, and what (if anything) I need from the client. Professional, plain English, under 180 words. Restate only what I give you — do not add advice, recommendations, or numbers I didn't provide. Discussion points and next steps: [PASTE — no client PII or specific holdings] ```

Why it works: the model recaps cleanly and warmly, and 'restate only what I give you' prevents it from adding a recommendation or a figure that wasn't actually discussed — which protects both accuracy and your compliance record. Adapt with the Business Email Generator.

When to use: condensing a prospectus, fact sheet, or disclosure into the points that matter, with sources to verify.

``` Summarize the document below for my own preparation. For each point, cite the page or section and quote any critical figure (fee, minimum, risk factor) exactly. - One-sentence bottom line. - Key features, costs, and risks, each with its source location. - Anything ambiguous that I should confirm before discussing it. Work only from the text. Do not infer figures that aren't stated, do not characterize it as good or bad, and do not recommend it. Document: [PASTE] ```

Why it works: requiring a page/section citation and an exact quote for every fee and risk factor makes the summary verifiable against the source — essential when you'll rely on these specifics. 'Do not characterize or recommend' keeps it neutral so your judgment, not the model's, drives any decision.

When to use: a structured, educational first draft answering a common client question, which you make accurate and personal.

``` A client asked: [QUESTION]. Draft a general-education response for me to review, personalize, and verify before sending. - Explain the relevant concept in plain English. - Lay out the considerations on each side, neutrally. - List the specific facts about the client I'd need to actually advise (as [BRACKETS]) and note where the answer depends on their situation. Do NOT give a recommendation, predict outcomes, or assume facts about this client. Flag clearly that this is general information. ```

Why it works: forcing the model to list the client facts it would need — as brackets you fill — keeps it from inventing a personalized answer, and the neutral 'considerations on each side' framing gives the client real understanding without a steer you didn't sanction.

When to use: structuring a recurring review meeting so you cover what matters without scripting advice.

``` Build an agenda and talking points for a [annual / quarterly] client review meeting covering these themes: [LIST]. For each theme: the goal of discussing it, the questions to ask the client, and the general topics to cover. Sequence the agenda so it flows naturally and ends on next steps. Do not script any recommendation or use client-specific numbers. Keep everything as a discussion structure, not a decision. ```

Why it works: the model is strong at sequencing a meeting so each item builds toward a clear close, and keeping it to 'discussion structure, not a decision' means you bring the actual advice while the model handles the scaffolding and the questions.

When to use: rewriting something you've written so a client can read it without a finance dictionary.

``` Rewrite the text below in plain English for a client with no finance background. Keep every fact, figure, and caveat exactly as written — you are simplifying language only, not changing meaning or numbers. - Replace jargon with plain words (define a term only if unavoidable). - Shorten sentences; use a calm, reassuring-but-honest tone. - Preserve all risk language and disclaimers verbatim. If simplifying a sentence would change its meaning, leave it and flag it. Text: [PASTE — no client identifiers] ```

Why it works: 'keep every fact and figure exactly, simplify language only' is the constraint that lets the model improve readability without quietly altering a number or softening a required risk disclosure. The flag-don't-distort rule protects meaning where simplification is risky.

When to use: building a library of general, educational FAQ answers for your site or client portal, pre-compliance.

``` Draft general-education answers to these common client questions: [LIST]. These will be reviewed by compliance before publishing. For each: a clear, neutral, plain-English answer; the considerations involved; and a closing line that this is general information, not individual advice. Do NOT make recommendations, cite performance figures, or state firm- specific facts. Use [BRACKETS] for anything I must supply or verify. ```

Why it works: framing the output as pre-compliance drafts with built-in 'general information' disclaimers sets the right expectation, and the brackets-and-no-figures rule keeps the model from inventing the firm-specific or performance claims that trigger review problems. Build the page itself with the FAQ Section Generator.

When to use: before a piece goes to a client or compliance, to catch language that could read as advice or a guarantee.

``` Review the client communication below as a compliance-minded reader. Flag any sentence that: - could be read as a recommendation, forecast, or guarantee; - promises or implies a specific outcome or return; - states a figure or fact that should be sourced or verified; - is unclear, jargon-heavy, or could confuse a client. For each flag, quote the sentence and suggest neutral alternative wording. Do not rewrite the whole thing — just flag and suggest. Communication: [PASTE] ```

Why it works: the model is a useful first-pass compliance reader for spotting outcome language and unsourced claims before a human reviewer does. Asking it to quote-and-suggest rather than rewrite keeps you in control of the final wording while surfacing exactly what to fix.

The most important rule for advisors using AI: the model is not a compliance officer, a source of facts, or an advisor. It fabricates performance figures, fees, and product features that look entirely real, and it can produce language that reads as a recommendation or a guarantee. None of the prompts above ask it to recommend, forecast, or supply specific figures — they keep it drafting and explaining from what you provide. Every client-facing piece remains your communication: fact-check it, confirm it fits the client's situation and your suitability obligations, and run it through your firm's supervision, advertising, and recordkeeping process.

Choosing a model: for prep, explanation, and drafting, any current frontier model works well. As of June 2026, Claude Opus 4.8 ($5 in / $25 out per 1M, with a 1M-token context window for long disclosures) and Gemini 3.1 Pro (~$2.00 / $12.00) both handle large documents; gpt-5.5 ($5 / $30) is a strong general option. None is a source of advice. Verify rates on each provider's live page (OpenAI, Anthropic, Gemini).

Sources and further reading: DAIR.ai Prompt Engineering Guide, Learn Prompting, OWASP LLM Top 10 (2025), Claude prompt engineering overview. This article is general information, not investment, legal, or tax advice. Pricing current as of June 2026.