Most brokers do not lose accounts to a cheaper quote. Instead, they lose them to silence. The renewal date passes, nobody called, and the client took the first offer that showed up in their inbox.
These Claude prompts for insurance brokers are built for that problem. Below are 25 prompts covering the work that keeps accounts: qualifying leads, comparing policies, running renewals, updating claims, handling objections, spotting cross sells, and writing emails that stay compliant.
A note before you start. These prompts produce drafts, not advice. Carrier rules, state regulations, and licensing requirements vary, so a licensed broker must review every output before it reaches a client. Never send a coverage or premium statement you have not verified against the actual policy.
Qualify every inbound quote request automatically.
Why brokers lose renewals they already had
A book of business decays quietly. Because nothing breaks, nothing signals a problem until the non renewal notice arrives.
Three patterns cause most of it:
- No contact between the sale and the renewal. Twelve months of silence, then a call asking for money.
- Renewal outreach that starts too late. Two weeks out, the client has already been shopped.
- Claims handled without updates. The client felt alone during the one moment that mattered.
Each is a communication failure, not a pricing failure. Therefore each is fixable with better throughput, which is exactly what these Claude prompts for insurance brokers provide.
Worth naming the scale of it. A broker managing 400 personal lines accounts cannot personally touch each one four times a year. That is roughly 1,600 conversations. So the work either gets systematized or it silently does not happen.
Why Claude for insurance work
Claude suits this work for two practical reasons.
It handles long documents well. A commercial policy runs dozens of pages. Because Claude keeps long context coherent, you can paste a full declarations page or a policy section and ask for a plain language summary without it losing the thread.
Its default tone is careful. Insurance writing has to be accurate and measured. Claude tends to hedge rather than overclaim, which is the right instinct when a sentence could be read as a coverage promise.
That said, the same rule applies as with any model. It drafts, you verify. If you prefer a different tool, our ChatGPT prompts for insurance brokers cover the same ground.
25 Claude prompts for insurance brokers
All of these Claude prompts for insurance brokers follow the same pattern. Replace the bracketed fields with your own details. In addition, append a tone instruction so the output sounds like you rather than a generic assistant.
Lead qualification (prompts 1 to 4)
1. “You are helping an insurance broker qualify an inbound lead for [LINE OF BUSINESS]. Write 10 screening questions that establish need, current coverage, renewal date, decision maker, and timeline. Conversational, not an interrogation.”
2. “Turn these messy call notes into a structured lead record with fields: contact, line of business, current carrier, renewal date, stated pain, and next step. Notes: [PASTE NOTES]”
3. “From this lead record, list what is still missing before I can quote accurately. Output as a short follow up checklist. Record: [PASTE RECORD]”
4. “Draft three short qualifying questions I can ask in the first 60 seconds of a call to find out whether this prospect is shopping seriously or just collecting numbers.”
Policy comparison summaries (prompts 5 to 9)
5. “Summarize this declarations page into plain language for a client: what is covered, key limits, deductibles, and notable exclusions. Do not interpret coverage for any specific scenario. Page: [PASTE REDACTED DEC PAGE]”
6. “Build a side by side comparison table of these two quotes, rows for limits, deductibles, endorsements, and premium. Note differences neutrally. Quotes: [PASTE A] and [PASTE B]”
7. “Rewrite this policy section at a ninth grade reading level, keeping every condition and limit exactly as written: [PASTE SECTION]”
8. “List the questions a client should ask before choosing between these two options, framed neutrally. Comparison: [PASTE COMPARISON]”
9. “Draft a one page summary explaining what changed between last year’s policy and this year’s renewal. Changes: [PASTE CHANGES]”
Renewal outreach (prompts 10 to 14)
10. “Write a renewal outreach email to send 90 days before expiration. Warm, brief, and focused on reviewing whether the coverage still fits. No premium claims. Under 140 words.”
11. “Write the 60 day follow up for a client who did not reply to the 90 day email. Slightly more direct, still no pressure.”
12. “Draft a 30 day renewal call script with an opening, three review questions, and a clear next step. Include a branch for ‘I am already shopping this.’”
13. “Write a short message to a client whose renewal premium increased, explaining that market conditions affect rates and offering a coverage review. Do not promise savings.”
14. “Create a renewal sequence outline across 90, 60, 30, and 7 days, listing the channel and the goal of each touch. Output as a table.”
Claims status updates (prompts 15 to 18)
15. “Write a client email confirming a claim was filed, what happens next, and roughly what the process involves. Reassuring, no timeline promises.”
16. “Draft a status update email for a claim that is still pending with no new information. Acknowledge the frustration honestly. Under 120 words.”
17. “Rewrite this adjuster message into plain language a client can understand, preserving every fact and condition: [PASTE MESSAGE]”
18. “Draft a checklist a client can follow after a [CLAIM TYPE] loss, covering documentation, immediate steps, and what to keep.”
Objection handling (prompts 19 to 21)
19. “I sell [LINE OF BUSINESS]. Write empathetic, non pushy responses to these objections: it is too expensive, I already have coverage, I need to think about it, my brother in law does insurance. Two sentences each.”
20. “Rewrite this response so it acknowledges the concern before answering it: [PASTE RESPONSE]”
21. “Draft three questions I can ask when a prospect says the price is too high, designed to uncover what they are actually comparing.”
Cross sell suggestions (prompts 22 to 23)
22. “Given this client profile, list coverage gaps commonly discussed for a similar profile, framed as review questions rather than recommendations. Profile: [PASTE PROFILE]”
23. “Write a short, non salesy message offering an annual coverage review to a client who currently holds only one policy with us.”
Compliance safe client emails (prompts 24 to 25)
24. “Review this client email and flag any sentence that could read as a guarantee of coverage, a promise of savings, or advice beyond a licensed broker’s role. Do not rewrite, just flag. Email: [PASTE EMAIL]”
25. “Rewrite this marketing message to remove any implied guarantee while keeping it warm and specific: [PASTE MESSAGE]”
Try these on your own renewal list.
Feeding Claude a policy document without leaking client PII
The prompts above assume redacted input, so here is the workflow that makes redaction quick.
Build a substitution habit before you paste. Client name becomes “the insured”, policy and claim numbers become [REDACTED], and addresses become “the insured location”. Because the model reasons from structure, none of those details improve the draft.
Then reverse the substitution in your document editor on the way out. The identifying data never reaches the model at all.
Two more habits matter:
- Do not paste raw PDFs. Extract the text you need, redact it, then paste that.
- Keep one chat per task, not per client. A chat about “renewal email structure” holds no client data, so it stays reusable.
Finally, use a business tier account where chats are excluded from training and retention is controlled.
Building a renewal sequence from these prompts
A single email is a task. However, a sequence is a system, and only the system protects your book. This is where Claude prompts for insurance brokers stop being a novelty and start being infrastructure.
Start with prompt 14 to draft the skeleton. Then use prompts 10 through 13 to write each touch. Load them into your CRM as templates with merge fields.
The result is that renewals stop depending on whoever remembers. Instead, every account gets four touches whether or not it was a big one. As a result, the accounts you quietly lost each year stay.
One practical note on sequencing. Send the 90 day email before the renewal quote arrives, not after. Because the conversation then starts with coverage rather than price, you are reviewing a relationship instead of defending a number.
Catching the quote calls that come in while you are with a client
Here is the gap prompts cannot close. Every prompt above works on information you already have. None of them answer the phone.
That matters because inbound quote requests are perishable. Someone shopping auto insurance is calling three agencies in one sitting. Because you are on another call, the second and third agency get the conversation.
Dapta closes that gap with AI voice agents that answer every call, qualify the caller, and hand your team a structured summary.
What an AI agent does on an insurance call
- Answers every inbound call, including evenings and weekends
- Asks your qualifying questions, the same ones in prompt 1
- Speaks natural English and Spanish, with regional accents that sound local
- Books the quote appointment on your calendar during the call
- Writes the lead record to your CRM, so nothing is retyped
- Stays inside your compliance language, because it reads from your script
The pairing works cleanly. Dapta captures the conversation, and then your Claude prompts for insurance brokers turn that capture into a comparison summary and a follow up email.
For the full picture, explore our related guides:
- Claude prompts for sales teams, the parent guide this list builds on
- ChatGPT prompts for insurance brokers, the same tasks with a different model
- Best AI tools for insurance agencies, a wider tooling roundup
- AI for insurance agents, the practice level overview
- AI lead generation for insurance agencies, filling the top of the funnel
- AI answering service for insurance agencies, for agencies missing calls
- More AI prompt libraries, the full collection
Frequently asked questions
Is it compliant to use Claude for client communication?
It can be, with review. These prompts draft language that a licensed broker then checks against the actual policy and applicable state rules. Because compliance obligations sit with you and not the tool, never send AI generated coverage statements unverified.
Can Claude tell me whether something is covered?
No. Coverage depends on the specific policy form, endorsements, exclusions, and the facts of the loss. Claude can summarize what a document says, but the coverage determination stays with you and the carrier.
Is it safe to paste a policy into Claude?
Only after redaction. Remove names, policy numbers, and addresses first. In addition, use a business tier account where your conversations are excluded from training.
Will clients notice the emails were AI drafted?
Not if you direct the tone and edit before sending. Add an instruction such as “warm, plain language, under 140 words, no jargon.” For example, paste one of your own past emails and ask Claude to match that voice.
Does this replace my service team?
No. These Claude prompts for insurance brokers remove drafting time so your team spends more of the day talking to clients. A person still reviews everything that goes out.
Where should I start?
Start with the renewal sequence, prompts 10 through 14. Because renewals are predictable and already on your calendar, that is where a small system change produces the clearest return.
Stop losing quote calls to voicemail. , no card required.
