Automation Recipe: Scale Year-End Tax Planning Letters Across Your Client Base

What This Builds

A semi-automated system for generating personalized year-end tax planning letters for every client in your book — each tailored to their specific situation, entity type, and planning opportunities. Instead of writing 80 planning letters one at a time in November, you build a client data template, run it through a structured AI workflow, and produce all 80 drafts in an afternoon. You review and send; Claude drafts.

Prerequisites

• Claude Pro at claude.ai ($20/month — needed for longer context to handle client data)
• Excel or Google Sheets (any version)
• A client list with basic planning facts (you'll build this list as part of the setup)
• 2–3 hours for initial build; 1–2 hours per year thereafter to refresh client data

The Concept

Instead of writing each letter from scratch, you build a client data template — a spreadsheet with the key facts that drive planning recommendations for each client. Then you run batches through a structured Claude prompt that knows your firm's style and the current year's planning priorities. Claude produces a personalized draft for each client; you review and personalize further before sending. Think of it as turning a week of letter writing into an afternoon of quality review.

Build It Step by Step

Part 1: Build Your Client Data Template

Create a spreadsheet with one row per client. You don't need every piece of tax data — just the facts that drive year-end planning conversations:

Columns to include:

Client Name | Client Type | Entity | Current Year Income (est) | Last Year Tax Owed/Refund | Estimated Q4 Taxes Outstanding | Key Events This Year | Recommended Planning Actions | Notes/Special Situations

Fill in the key planning triggers for each client:

• Key Events This Year: "Sold business," "New rental property," "Salary increase," "Large capital gain," "Started retirement"
• Recommended Planning Actions: Pre-populate with your actual planning recommendations for each client (you'll review this anyway — this step forces you to think about each client)
• Current Year Income: Approximate — helps Claude calibrate the magnitude of recommendations

Time required: 30–60 minutes to complete for a 50-client book if you work from your tax software's client list.

Part 2: Create Your Master Planning Letter Prompt

This is the prompt that drives all letter generation. Customize once, use for all clients:

You are a professional tax advisor writing year-end planning letters for clients. You work at [FIRM NAME], a [firm type]. Our letters are professional, proactive, and advisory in tone — not overly formal or legalistic.

Current year planning priorities (customize these annually):
- Estimated tax payments: Q4 due January 15 — ensure clients who owe are reminded
- Retirement contributions: IRA contribution deadline April 15; 401k December 31
- Capital gains: opportunity to harvest losses to offset gains before year-end
- Required Minimum Distributions: clients over 73 must take RMDs by December 31
- Business deductions: equipment purchases for §179 / bonus depreciation before year-end
- Roth conversions: consider for clients in temporarily lower bracket years
- Charitable giving: donor-advised fund contributions by December 31 for current year deduction

Write a year-end planning letter for the following client. Personalize to their specific situation — only mention relevant planning items, not all of them. Under 350 words. Professional but warm. End with a call to action to schedule a planning call.

Client data:
[I will paste one client's data row here]

Part 3: Process Clients in Batches

Open Claude Pro. Paste your master prompt at the start of the conversation. Then paste one client's data and request their letter:

"Write the year-end planning letter for this client: [paste their data row from your spreadsheet]"

For each subsequent client in the same conversation, just paste their data row and say "Write the planning letter for this client." Claude maintains the prompt context throughout the conversation.

Processing speed: An experienced user can work through 8–10 client letters per hour — review each draft, make any adjustments, paste into your email system or letterhead template.

Part 4: Batch Processing Workflow (Optional Advanced Setup)

For larger books (100+ clients), consider a more structured approach:

1. Segment your client list by type (individual high-income, S-corp owners, retirees, real estate investors) — each segment gets a slightly different master prompt emphasizing their relevant planning items
2. Open one Claude conversation per segment — paste the segment-specific master prompt, then work through all clients in that segment
3. Review and personalize each draft before pasting into your email system — look for: accuracy of the planning recommendations, appropriate tone for this client, any client-specific nuances the AI missed

Part 5: Build a Template Letter Library

After your first year-end run, save the best-performing letter drafts (the ones you edited least before sending) as examples. Add them to your Claude Project (Level 4, Guide 1) under the Files tab. Future planning letters will match these quality examples.

Real Example: 15-Client Year-End Run

Setup: 15 business owner clients. You've filled in their data rows in your spreadsheet.

Client 1 data:

Name: Johnson Manufacturing | Type: Business client | Entity: S-corp | Income: ~$420K | Last Year: Owed $38K | Q4 Status: No estimated payment made | Key Events: Purchased $180K equipment October | Planning Actions: Ensure §179 election, consider bonus depreciation, Q4 estimated payment, maximize SEP-IRA contribution before filing deadline | Notes: Growing rapidly, may want to revisit compensation structure

What you type to Claude:

Write the year-end planning letter for this client: [paste data row]

What you get: A 280-word letter specifically mentioning the §179 election opportunity for the equipment purchase (with a brief plain-language explanation), reminding them about the Q4 estimated payment due January 15, and noting the SEP-IRA opportunity for additional tax savings. Ends with a prompt to schedule a call before December 31.

What you do: Read the letter, verify the §179 deduction limit is accurately stated, add one sentence about their growth trajectory ("As you continue to grow, this is also a good time to review whether your current entity structure and compensation package are still optimally structured"), and paste into your firm email template.

Time: 8 minutes per client vs. 30–45 minutes writing from scratch.

Total time for 15 clients: 2 hours vs. 7–11 hours.

What to Do When It Breaks

• Letters are too generic → Your client data rows need more specifics. "High income" produces vague output; "$420K S-corp income, bought $180K equipment, no Q4 payment" produces a targeted letter.
• Letters mention irrelevant planning items → In your master prompt, emphasize "ONLY mention planning items directly relevant to this client's data. Do not mention RMDs if they're under 70. Do not mention capital gains harvesting if they have no investment income."
• Tone doesn't match your firm's style → Add 2-3 sentences to your master prompt describing tone: "Our letters sound like they're written by a trusted advisor who knows this client personally — warm but professional, not corporate-speak."

Variations

• Simpler version: Skip the spreadsheet entirely. Just describe each client verbally in your prompt. Less systematic but still saves 50–60% of writing time.
• Extended version: Fully automate using Claude API + a simple script to loop through your spreadsheet rows and call the API for each client. Produces all drafts in 15 minutes. Requires basic programming knowledge or a technical staff member.

What to Do Next

• October: Build your client data spreadsheet while you still remember each client's year well
• November: Run your first batch — plan for 1 week to draft + 1 week to review and send
• December: Refine your master prompt based on which letters needed the most editing — save the best examples for next year

Advanced guide for tax manager professionals. These techniques use more sophisticated AI features that may require paid subscriptions.