Custom GPT: Your Life Insurance Product Knowledge Assistant

What This Builds

You'll build a Custom GPT loaded with your carrier product guides, underwriting cheat sheets, and product comparison summaries. Instead of searching PDFs or calling your carrier helpdesk during case prep, you ask your Custom GPT: "Which of my carriers offers the best rate for a 50-year-old male with controlled diabetes?" and get a structured answer in seconds. This is a product knowledge assistant that works 24/7 and knows everything you've uploaded to it.

Prerequisites

• ChatGPT Plus account ({{tool:ChatGPT.plan}} — {{tool:ChatGPT.price}}) — Custom GPT creation requires a paid plan
• 3-5 carrier product guides or rate tables in PDF format (download from carrier portals)
• A list of your top underwriting niches or health class rules if you have them documented
• Cost: {{tool:ChatGPT.price}}

The Concept

A Custom GPT is like a specialized colleague who has read every product manual you've ever received and can answer questions instantly. You load it with your specific carriers and products once — it never forgets, never puts you on hold, and is available while you're on a client call. Think of it as your carrier helpdesk that knows your specific product lineup.

Build It Step by Step

Part 1: Gather your source documents

Before you start building, collect these documents (PDFs work best):

• Rate tables or product summaries for your top 3-5 carriers
• Underwriting guideline summaries for each carrier (available from your IMO or carrier portal)
• A list of your common cases with any underwriting niches you've learned
• Any product comparison documents you've built or received

Tip: You don't need perfect, comprehensive documents. Even a rough Google Doc with your underwriting notes and a few carrier PDFs is enough to get started.

Part 2: Create your Custom GPT

1. Log into ChatGPT Plus at {{tool:ChatGPT.url}}
2. Click Explore GPTs in the left sidebar
3. Click Create a GPT (top-right corner)
4. You'll see the GPT Builder — a split screen with a configuration panel on the left and a preview on the right

Part 3: Write the GPT instructions

In the Configure tab (click it in the builder), fill in:

Name: "Product Knowledge Assistant — [Your Name]"

Description: "Helps me research carrier products, underwriting guidelines, and case placement"

Instructions (paste and customize):

You are a life insurance product knowledge assistant for [Agent Name], an independent life insurance agent. You have access to uploaded carrier product documents and underwriting guidelines.

Your job is to help the agent:
1. Compare carrier products for specific client situations
2. Look up underwriting guidelines for health conditions
3. Identify which carriers have favorable appetites for specific risks
4. Explain product features accurately and clearly
5. Help prepare for case placement conversations

Always be accurate. If you're not sure about something, say so — don't guess about underwriting decisions. Reference the specific document and section when possible.

Carrier lineup the agent works with: [list your carriers]
Product focus: [your main products — term, whole life, final expense, etc.]

Part 4: Upload your documents

1. In the Configure tab, scroll down to Knowledge
2. Click Upload files
3. Upload your carrier PDFs, rate tables, and underwriting documents (up to 20 files)
4. Give each file a descriptive name before uploading ("Carrier-A-Term-Product-Guide-2025.pdf")

What you should see: Each file appears in the Knowledge section as an uploaded document. The GPT will be able to search and reference these in its answers.

Part 5: Test and refine

Click Preview (the right side of the builder) and test with real scenarios:

1. "Which of my carriers has the best non-smoker preferred rate for a 45-year-old female, $500K 20-year term?"
2. "What's [Carrier A]'s underwriting stance on controlled Type 2 diabetes?"
3. "Compare the key differences between [Carrier A] and [Carrier B] for final expense coverage."

What good output looks like: Specific answers that reference the uploaded documents, not generic insurance knowledge. If it can't find specific data, it says so rather than guessing.

If outputs are too generic: Go back to Instructions and add: "Always search the uploaded documents first before drawing on general knowledge. If you can't find specific information in the documents, tell me that rather than answering from general knowledge."

Part 6: Save and access your Custom GPT

1. Click Save (top-right)
2. Set visibility to Only me — this keeps your carrier data private
3. Access it anytime from the GPT panel in the left sidebar under your account

Real Example: Case Placement Research

Setup: Agent works with 6 carriers, has uploaded product guides and underwriting summaries for each.

Scenario: Client is a 52-year-old male, mild sleep apnea (CPAP-compliant), on medication for mild anxiety, applying for $750K 20-year term.

Input: "Which of my carriers is likely to offer the best rate class for a 52-year-old male with CPAP-compliant sleep apnea and mild anxiety on medication, applying for $750K 20-year term? Is there any reason to prefer one carrier over another?"

Output: Specific carrier comparison based on uploaded guidelines — Carrier A's sleep apnea table shows preferred plus is achievable with CPAP compliance; Carrier B is more conservative; anxiety may be table rated at some carriers but standard at others. Agent goes to the right carrier first.

Time saved: 20-30 minute research session compressed to 2 minutes.

What to Do When It Breaks

• Answers are too generic (not using your documents) → In the Instructions, add: "Always reference the uploaded documents. If the answer isn't in the documents, explicitly say the documents don't cover this topic."
• Wrong information from old documents → Delete outdated files and upload current versions; carrier rates change regularly
• GPT can't find specific data → Your uploaded documents may not contain that specific information — supplement with a direct carrier call or updated document upload
• Document format not being read correctly → Try converting PDFs to plain text first; some PDFs (especially scanned forms) aren't machine-readable

Variations

• Simpler version: Don't upload documents — use the GPT Instructions alone to set up an insurance-context assistant that you query for general product knowledge questions
• Extended version: Add a "Carrier Appetite Quick Reference" document you maintain yourself — updated notes on which carriers are currently competitive for which types of cases

What to Do Next

• This week: Build the GPT with your top 3 carriers' product documents; test with 5 real case scenarios from your pipeline
• This month: Add more carrier documents and refine the Instructions based on where the GPT is giving weak answers
• Advanced: Share the concept with your IMO or upline — they may have more comprehensive carrier documents to share

Advanced guide for life insurance agents. Custom GPT requires a paid ChatGPT subscription. Do not upload client personal information or anything subject to privacy restrictions.