Skip to main content

What you can do with the Zocks MCP

What you get out of Zocks MCP once it's connected — tax planning, plan audits, relationship reads, referrals, compliance, coaching, pipeline and content — with ready-to-use prompts. Works the same in Claude, ChatGPT, or Copilot.

The Zocks MCP Server lets the AI tool you already use read your Zocks meeting intelligence on demand. This article is about what to do with it once it's connected: the questions worth asking, the work it takes off your plate, and the outputs you can hand to a client, a colleague, or compliance.

Everything below works the same whether you connected Claude, ChatGPT, Microsoft Copilot, or your own in-house AI. The prompts are plain language, so they carry across tools. Two things differ: how you bring Zocks into a chat (in ChatGPT you type @Zocks; in Claude and Copilot you just ask), and pre-built Default Prompts, which are available in Claude today.

Every prompt below is a starting point. Change the names, the timeframe, and the output format and it still works.


Why the answers hold up

The value isn't that an AI tool can write about financial planning. It's that this one answers from your own client conversations.

  • Grounded in what was actually said. Answers come from your transcripts, summaries, AI results, conversation insights, and contact and household records. Ask for citations and you get the meeting date and the client's own words back.

  • Read-only. Your AI tool can read Zocks data. It cannot create, edit, or delete anything in Zocks.

  • Your permissions apply. It sees the meetings you can already see in Zocks, and nothing else.

  • Logged. Every MCP tool call is recorded: which user, which AI tool, when, and what data was accessed.

  • Gaps get flagged, not filled. Ask it to tell you what data was missing rather than work around it, and you'll know how much weight to put on the answer.

  • Still review it. Working from your real data lowers the risk of a made-up answer. It doesn't remove it.

What MCP is for, and what to use instead

MCP is built for depth on a defined subject: one client, one household, one advisor, your recent meetings, or your firm's aggregated conversation insights. Questions that require reading every transcript across your whole book are a different job.

If you're asking about

Use

One client or household — their full meeting history

Zocks MCP

Your own recent meetings — last week, last five, last ten

Zocks MCP

One advisor's meetings — coaching, scoring, talk track

Zocks MCP

Firm-wide patterns — topics, client outlook, sentiment, pipeline

Zocks MCP, via your Conversation Insights

Every client across your entire book of business

A quick lookup while you're working inside Zocks

The short rule: a client, a household, an advisor, or your last few meetings → MCP. Your entire book → Client Queries.


So questions like "find every meeting where we discussed alternative investments this quarter", "which clients discussed life insurance but don't have a policy", or "show me five ways fee objections were handled across the firm" aren't MCP questions. They need thousands of transcripts. Take those to Client Queries.


What to bring with you

MCP supplies the conversation side: what the client said, when, and how it changed. Several of the highest-value use cases also need a document or a dataset. Bring it and the output gets sharp; skip it and the output gets generic.

For this

Bring

Tax planning

The client's tax return. This one is non-negotiable — without it the analysis is generic and can mislead. Either through the Holistiplan integration or by attaching it in your AI tool.

Plan audits and scenario work

The current plan or a plan export from your planning tool, such as eMoney or Conquest.

Harvesting and gain realization

Holdings-level data with cost basis and holding period, from Orion, Black Diamond, or a report you attach.

Meeting scoring

Your firm's own scorecard. Your dimensions and your scale take precedence over any defaults.

Letters, proposals, reports

Your letterhead or report template, so formatting stays consistent.

Multi-entity estate work

Trusts, business entities, and ownership relationships mapped in your CRM or estate platform first. Cross-entity reasoning needs the structure to exist somewhere.

Note: documents you attach go into your AI tool, not into Zocks, so they're governed by your AI tool's plan and retention settings. See Zocks MCP & Claude data privacy.

One client, one household

This is where MCP does its best work. Everything you know about a family is already in the recordings; it's just spread across two years of meetings and nobody has time to re-read them.

Tax planning

Tax planning is one of the highest-impact things you can do for a client, and one of the slowest. The facts that matter are usually already in your meetings: an equity grant mentioned in passing, a rental property, a business sale two years out, a parent's estate. Pulling them together by hand is hours of reading, and tax season is exactly when you don't have those hours.

Ask:

  • "What are the best opportunities to minimize the Morgans' taxes this year and over the next five years? Use their tax return, their plan, and everything they've told us in meetings."

  • "Split those into current-year actions with a deadline and multi-year strategies. Rank by estimated impact and cite the meeting each one came from."

  • "What are the tax questions nobody has asked this household yet?"

  • "Build a CPA handoff: what changed this year and what their accountant needs before filing."

What you get back: a prioritized list across both horizons — contribution gaps, bracket positioning, elections with deadlines, then conversion windows, drawdown sequencing, business structure, and planned transactions — each cited to the meeting and the client language that surfaced it, with data gaps flagged rather than guessed. Anything needing a CPA or attorney is called out. In Claude, Tax Opportunities runs a version of this in one click.

Plan audits: gaps, contradictions, and staleness

A plan gets built once, and then the client keeps talking. Six months later, what they said in a meeting and what the plan assumes have quietly drifted apart. Checking one against the other is precise, valuable work that nobody has time for.

Ask:

  • "Run a full audit of the Wilsons' financial plan against what they've told us in meetings and flag every gap and contradiction."

  • "The Nguyens' life has changed a lot in the last 18 months. How outdated is their plan?"

  • "Which plan assumptions haven't been confirmed with them in the last twelve months?"

  • "What happens to the Robinsons' plan if Maria doesn't go back to work after maternity leave?"

What you get back: a numbered gap and contradiction list you can walk into the next review with — retirement age stated in a meeting versus modelled in the plan, spending target versus described lifestyle, missing long-term care provision, a business valuation absent from net worth, a claiming age that doesn't match what they said. Life events detected from meetings get compared against the plan systematically, so nothing depends on you remembering it.

Everything they've said about one topic

Every household has a subject that has come up four times and never landed. Whether it stalled on your side or theirs is hard to reconstruct from notes.

Ask:

  • "Give me everything discussed with the Johnson family in the last six months about college planning."

  • "What has David Chen said about his business across all our meetings?"

  • "How has Sarah Thompson's risk tolerance changed over the last two years? Distinguish what she says from how she's actually behaved."

  • "What did I commit to in my last three meetings with this household that I haven't delivered?"

What you get back: a chronological view with a date against every statement — how the position evolved, decisions and commitments made, open action items raised and never resolved, and contradictions between meetings. Where stated tolerance and actual behaviour diverge, that divergence is itself the finding, and you'll get a flag if a formal re-assessment is worth doing before the next proposal.

The person, not just the account

What a client is afraid of, who in the family actually decides, and whether they're warming up or cooling off are all in the recordings. They just aren't in a field anywhere.

Ask:

  • "What are the core motivators and emotional drivers for this household across all our meetings? Quote them."

  • "Using only what James Harrington has told me, write a one-page personal brief — family, values, fears, goals, hobbies, how he talks about money — so someone meeting him for the first time would feel like they've known him for years."

  • "Based on all my meetings with the Hendersons, how engaged are they, and are there any warning signs?"

  • "Has sentiment been improving or declining, and what was being discussed when it shifted?"

  • "Predict the three to five themes most likely to come up in our next conversation, and give me a proactive talking point for each."

What you get back: motivators ranked by recurrence and emotional weight, with the meeting each first appeared in and whether it's ever been formally addressed. On the retention side: meeting frequency trend, who has stopped speaking, unprompted fee questions, self-directed research mentions, each evidenced with a date, plus a retention read and a next step framed for that specific client rather than a generic check-in. In Claude: Client Motivators and Sentiment Trend.

Planning coverage gaps

It's easy to cover the same three topics with a household for years and never notice the ones you've skipped.

Ask: "Are there any financial planning topics I've never covered with this household across all our meetings?"

What you get back: every core planning area — tax, estate, insurance and risk, education, equity comp, investment management, retirement forecasting, cash flow and debt, goals and profile — marked as discussed in detail, mentioned in passing, or never addressed, with the meetings that covered each and a prioritized agenda for the next one. In Claude: Planning Topic Gaps.

The wider household: family, beneficiaries, and their other professionals

Clients mention their brother's business sale, their daughter's divorce, the attorney who drafted the trust, and the accountant they've used for twenty years. Those names carry real planning weight and they're scattered across years of conversation.

Ask:

  • "Identify every beneficiary mentioned across all our meetings with Dorothy Simmons, with any context about the relationship — estrangements, life changes, complicated dynamics — that suggests a designation may no longer reflect her wishes."

  • "Map out this family: who's who, who's involved in decisions, and which family members we've never actually met."

  • "Who are this household's outside professionals, and what has each of them been involved in?"

  • "Analyze the Johnson household across all related entities and tell me if there are estate, tax, or structural optimization opportunities."

  • "Based on everything the Garcias have said, who in their network is most likely to need an advisor right now?"

What you get back: for referrals, a ranked list of named individuals with the signal behind each — a business sale, a divorce, a promotion, a windfall, no advisor mentioned — the meeting it came from, and a natural conversation opener built on what your client actually said about them, not a generic referral script. For beneficiaries, a plain-language explainer of what a policy or account means for each person named.

Prospects and referrals

The prospect side has its own set, and it's some of the fastest value to demonstrate.

Ask:

  • "I'm meeting Kevin Walsh for the first time. He was referred by Paul and Anne Murphy. Scan all my meetings with the Murphys and pull everything they've mentioned about Kevin — shared history, mutual connections, schools, hometowns, hobbies, family overlap. Give me conversation hooks I can reference naturally."

  • "Review all my meetings with David Chen and tell me exactly where things stalled — the last unresolved objection, any commitment that never got followed up, and the topics that clearly energized him. Then draft a short, warm re-engagement note built around one of them."

  • "Across my prospect and first meetings from the past two weeks, what objections came up most? For each, show what language worked in follow-ups, and flag any objection I've never successfully addressed."

What you get back: a rapport cheat sheet before a first meeting, a specific diagnosis of where a stalled relationship stopped moving, and an objection playbook drawn from your own successful conversations rather than a sales book.

Deciding your next move

Sometimes you don't want a report. You want to know what to do next and why.

Ask:

  • "For this household, what's the single highest-value thing I could do next? Show me the runner-up options you considered and why you ranked them lower."

  • "What have they raised more than once that we've never acted on?"

  • "What's the most important question nobody has asked this household yet?"

  • "Who outside our firm has to act for this to happen — their CPA, their attorney, an HR department?"

What you get back: a recommendation with the reasoning shown, which matters more than the recommendation itself. You can disagree with a ranking when you can see what it was ranked against.

Compliance and documentation

Ask:

  • "Generate a compliant meeting summary for yesterday's Johnson meeting that I can log in the CRM."

  • "Flag anything said in my last five meetings with the Garcias that could create a compliance or suitability concern."

What you get back: for summaries, a CRM-ready record — date, attendees, purpose, topics, information the client shared, recommendations with appropriate framing, action items with owner and due date, sentiment note, next meeting — every item sourced from the transcript with nothing invented. For scans, a numbered list of flagged items with the meeting date, the exact language, the concern type, and a recommended review action. None of it is presented as a confirmed violation and nothing gets dismissed for you; it surfaces for you and compliance to review.

Client-ready deliverables

Ask for a format and you get a document you can send, not a chat reply you have to rewrite.

Ask:

  • "Using only language and priorities Maria Gonzalez has expressed in our meetings, draft a one-page narrative proposal for a Roth conversion strategy. Use her own words wherever possible — it should feel written for her, not from a template."

  • "Based on everything George and Carol have shared about their retirement accounts, write them a personalized first-year RMD guide: when the first distribution is due, the key deadlines and decisions, how it interacts with their tax situation as we've discussed it, common first-year mistakes, and a simple checklist. Plain language they can read themselves."

  • "Write the Morgans' annual letter from the last twelve months of meetings — warm, in my voice — and flag anything you weren't sure about."

  • "Prepare this household's annual review: goals then versus now, what changed in their life and their finances, decisions taken, commitments made, and what each of their outside professionals needs to know afterwards."

  • "Based on personal details Tom and Linda have mentioned — hobbies, family, upcoming trips, things they're proud of — suggest five thoughtful gift ideas with a short handwritten note for each."

What you get back: a report, letter, PDF, document, or spreadsheet in the format you asked for. Attach your letterhead or template to keep it consistent. In Claude: Annual Client Letter.

Your own recent meetings

The other natural scope for MCP is your last week or your last ten meetings — small enough to read properly, wide enough to show a pattern.

Ask:

  • "Review all my client meetings from the past week and tag each relationship as warming, cooling, or steady, with the specific moment behind each signal. Then give me a ranked call-back list, starting with anyone cooling."

  • "Look at my last five meetings and flag any life events — new job, retirement, inheritance, new baby, home purchase, business sale."

  • "Any outside money mentioned recently? Old 401(k)s, brokerage accounts elsewhere, a spouse's accounts, inheritance, sale proceeds. Group it by household and tell me what's new versus what's been raised before."

  • "Did anyone mention wanting to refer someone, or ask about a service we offer, without a follow-up from me?"

  • "Scan my meetings from the past two weeks and pull every instance where a competitor, another advisor, or an alternative product came up. Summarize what clients said and what concern each signals."

  • "Who am I seeing today, and what mood is each of them likely walking in with?"

What you get back: a short list of specific moments with the client, the meeting, and what was said, so you can decide what deserves a call. The competitor scan can go one step further — ask for a plain-language client-facing FAQ addressing the most common misconceptions, written to inform rather than defend, ready for marketing to polish.

Your firm and your team

Client work is the core of the business, but for a firm that's scaling, internal quality is just as important. Your firm-level conversation insights — topic frequency, who drove each topic, client outlook, sentiment, held-away mentions — are aggregated already, so MCP can reason across the whole team without reading every transcript.

Coaching and advisor performance

Most firms already know how they want a meeting to run. The constraint has been that nobody can sit in on every meeting.

Ask:

  • "Here's our meeting scorecard. Score this advisor's last meeting against it — a score, a rationale, the strongest moment, and the clearest missed opportunity for each dimension."

  • "Which advisors are proactively bringing up estate planning versus waiting for clients to raise it? Flag anyone consistently reactive."

  • "Which advisors haven't discussed tax-loss harvesting with any clients in the last three months? I want to flag it before year end."

  • "Which advisors are generating the most client questions on alternatives — is that engagement or an explanation gap?"

  • "How has Sarah's coverage of market volatility evolved over six months? More proactive, or still client-driven?"

  • "Give me a full audit of tracked topics across the firm this quarter — which is everyone covering, and which barely come up?"

  • "Create a sales presentation from Andrew's last five discovery meetings and his talk track."

What you get back: scoring against your criteria rather than generic ones, with a compliance flag section and a recommended focus for the next meeting. At team level, ranked advisor lists with coaching flags, before-and-after trend narratives, and a coverage picture across the firm. The talk track builder pulls a strong advisor's actual questions and transitions verbatim into something the rest of the team can learn from. This is how coaching becomes measurable rather than anecdotal. In Claude: Evaluate Last Meeting, with your scorecard attached.

Business development and pipeline

Ask:

  • "Summarize the held-away asset pipeline across the firm this quarter. Which advisors have moved the most mentions to client agreement, and where is progress stalling?"

  • "Are there investment products that come up a lot in meetings but never generate new opportunities? I want to know where we're leaving money on the table."

  • "Who are my top three advisors at converting held-away mentions into agreements, and what are they doing differently?"

  • "Clients are frequently asking about income-generating strategies. Are advisors actually bringing up the relevant products, or are we missing the handoff?"

What you get back: pipeline by advisor and stage with stall points named, a high-discussion / low-conversion gap table, a conversion funnel comparison with qualitative observations about what the best advisors do, and alignment analysis between what clients are asking about and what's actually being recommended.

Content and client communication

Ask:

  • "Based on the most client-driven topics and questions from the last 30 days, draft a short client newsletter that proactively addresses what clients are asking about most."

  • "Where are our advisors generating the most client questions? Build me a 60-minute team training agenda that addresses the biggest confidence gaps."

  • "Top five topics clients raised this month, with brief talking points for each so advisors can get ahead of them this week."

  • "Suggest four webinar topics based on client questions and outlook over the past six months, with a rationale for each."

What you get back: a newsletter draft ready to review, a training agenda with each item traced back to the data behind it, a weekly talking-points brief, and a ranked event topic list. Content built from what your clients are actually asking, rather than what the industry is publishing.

Zocks alongside your other connectors and live web data

Your AI tool can pull from more than one source in a single question. This is where some of the strongest firm-level use cases live: your conversation data on one side, what's happening in the world on the other.

Ask:

  • "Pull our top discussion topics for the last 30 days. Then search financial media for the most prominent market topics this week. What's getting heavy coverage that barely shows up in our advisors' conversations? Flag the top three readiness gaps and how to brief the team."

  • "Pull our client outlook data and the topics generating the most client questions. For the topic showing the most anxiety, search for what analysts are actually saying. Are our clients' concerns well-founded or media-driven, and what should advisors be saying?"

  • "Search for regulatory and tax policy developments coming into effect this quarter. Which have low mention frequency in our conversations? Prioritize them and give me talking points."

  • "Based on what Alex Park has told me about his role at a biotech startup and his equity compensation, research the current state of the sector and flag anything recent that could affect the value or timing of his options. Suggest a conversation topic for each finding."

  • "Check my Zocks meetings from last week against my calendar this week. Do I have a follow-up booked for everyone I met with?"

What you get back: gap reports between the outside world and your firm's conversations, misconception briefs with talking points to correct the narrative, policy change summaries with the clients most likely to be affected, and emerging topics you can prepare for before clients start asking.

Worth knowing: MCP reads your Zocks data. It does not make live calls into your CRM, portfolio, or planning tools. Those systems connect to Zocks through the Zocks App Library, which is also where your admin enables MCP. If you want your AI tool to read one of them directly, connect it there as its own connector, and it can combine both in the same question.

Where to start

If you are

Start with

An advisor with a review next week

Planning coverage gaps, then the commitments you still owe them

Heading into tax season

Tax planning, run one household at a time, tax return in hand

Meeting a referral for the first time

The referral warm-up — two minutes of setup, immediately obvious value

A firm owner or team lead

Meeting scoring against your own scorecard, then the proactive-versus-reactive topic check

In operations or marketing

A client newsletter from this month's top client questions

In compliance

A compliance and suitability scan on recent meetings, plus the MCP audit log in Zocks

Getting better results

  • Name the client the way Zocks does. Use the contact or household name as it appears in your records. If several match, your AI tool will ask you to confirm before it runs.

  • Say what format you want. A PDF or document to share, a spreadsheet to work with further, a report to read, a letter to send.

  • Ask for citations. "Cite the meeting date for each point" makes the output reviewable and makes gaps in your data obvious.

  • Ask it to flag gaps rather than fill them. "Tell me what data was missing instead of estimating" is the single most useful line you can add.

  • Attach your own templates. Scorecards, letterheads, report formats. Your criteria beat generic ones every time.

  • Keep one chat or project per household. The context stays clean and follow-up questions get better.

  • Keep the scope tight. A household, an advisor, a fortnight. If a question needs your whole book, it belongs in Client Queries.

  • In Claude, set your common Zocks tools to "Always allow" in Connector settings so you aren't approving every query.

  • Review before anything leaves your desk. With the exception of client letters and other explicitly client-facing outputs, these results are advisor intelligence, not client-facing advice.

Related articles

Questions, or a use case you want help building? [email protected] or the in-app chat. Connecting your own in-house AI: [email protected].

Did this answer your question?