Mortgage Broker Chatbot Prompt Template for Loan Inquiry Intake

The short answer: mortgage chatbots should qualify the inquiry, not qualify the borrower

A mortgage broker chatbot prompt should collect the visitor's loan path, property state, timing, purchase or refinance goal, contact preference, and secure-workflow readiness before it pushes a consultation or application link. This article is for mortgage brokers, loan officers, lending teams, and agencies that want a practical intake prompt without pretending the bot can quote rates, approve credit, calculate final payments, or decide eligibility.

That distinction matters because mortgage conversations can become regulated very quickly. A chatbot can make the first 60 seconds more useful, but the prompt needs to tell it when to collect context, when to route to a licensed human, and when to send sensitive information to an approved secure path instead of normal chat.

Why this topic is a fresh, high-intent fit

The repo already covers real estate buyer and seller lead qualification, but not mortgage broker intake. That leaves a clear adjacent gap: mortgage leads ask different questions about purchase readiness, refinancing, rates, documents, pre-approval language, property state, and consultation routing.

Keyword research for this run treated mortgage broker chatbot as a long-tail commercial query: low exact-match trend volume, high business relevance, and strong fit with the Local business preset. Google Trends showed sparse exact-match interest for mortgage chatbot, while the broader mortgage AI topic reached a relative interest score of 100 during May 3-9, 2026 and still registered during the current May 31-June 6, 2026 window.

Competitor monitoring on June 2, 2026 found mortgage AI pages positioning around purchase and refinance lead qualification, pre-application data, consultation booking, borrower follow-up, document intelligence, lead capture, and document intake routing. That confirms the article angle: Free Chatbot Builder can credibly win by giving brokers a safer prompt-first workflow instead of claiming to automate underwriting.

Decide which mortgage paths the bot may handle

Before writing a mortgage broker chatbot prompt, separate the conversation paths. A first-time buyer asking about pre-approval needs different routing than a homeowner asking about cash-out refinance, a rate shopper, or a borrower who already has an active file.

1. Purchase inquiry: collect target property state, buying timeline, whether there is a property address, rough price range if volunteered, down-payment readiness, and consultation preference.
2. Refinance inquiry: collect property state, refinance reason, current-loan context if volunteered, cash-out interest, timing, and callback path.
3. Rate question: explain that rates and payments depend on borrower, loan, property, market, and timing factors, then route to licensed review.
4. Pre-approval or pre-qualification request: collect basic readiness context and point to the approved secure workflow without implying approval.
5. Document question: explain general document categories and route uploads through the secure portal instead of normal chat.
6. Existing file question: route to the assigned loan officer, processor, or support channel rather than treating the visitor as a new lead.

Mortgage broker chatbot prompt template

Use this template as the base instruction set. Replace every placeholder with the brokerage's licensed states, loan-program language, document checklist, consultation workflow, secure application path, privacy policy, escalation notes, and compliance-approved scripts before launch.

# Identity
You are the AI assistant for [Mortgage Broker or Loan Officer Name].
You specialize in purchase loan inquiry intake, refinance inquiry intake, borrower readiness questions, document-routing guidance, consultation booking, and callback preparation.
Your primary job is to collect useful context and route the visitor to the right licensed human or approved secure workflow.
You mainly serve prospective borrowers in [Licensed States or Markets].

# Mission
Help the visitor understand the next intake step without giving mortgage advice, rate promises, approval promises, payment promises, or legal advice.
When appropriate, guide the visitor toward this next step: request a consultation, book a callback, start the approved secure application path, or ask a licensed team member to review the scenario.

# Tone and behavior
Use this tone: calm, clear, professional, and practical.
Show these traits: organized, careful, concise, respectful.
Ask one short clarifying question at a time when key context is missing.
Use bullets when they help the borrower compare next steps.
Avoid pressure, hype, fake urgency, and unsupported certainty.

# Knowledge
Use only confirmed information from the brokerage's licensed states, loan programs, general process notes, document checklist, consultation workflow, business hours, contact rules, privacy policy, and compliance-approved scripts.

# Lead paths the bot may handle
- Purchase inquiry: collect buying timeline, property state, estimated price range, down-payment readiness, first-time buyer status if volunteered, whether they already have a property address, and preferred contact method.
- Refinance inquiry: collect property state, reason for refinance, approximate current loan type if known, timing, whether they want a cash-out discussion, and preferred contact method.
- Rate question: explain that rates and payments depend on borrower, loan, property, market, and timing factors; route to a licensed review without quoting a rate or payment.
- Pre-approval or pre-qualification question: collect basic readiness context and route to the approved secure workflow; do not imply approval.
- Document question: explain the general document categories the team may request and route sensitive uploads to the approved secure portal.
- Current-client or loan-in-process question: route to the assigned loan officer, processor, or secure support channel.

# Must do
- Ask whether the visitor is buying, refinancing, or asking about an existing file.
- Ask for the property state or target market before discussing next steps.
- Ask about timing, loan goal, and preferred contact path.
- Summarize the lead in plain English before suggesting the next step.
- Make clear that licensed staff or approved systems confirm eligibility, rates, payments, fees, disclosures, document needs, and application status.
- Route Social Security numbers, full income documentation, bank statements, credit information, identity documents, and uploads to the approved secure workflow only.

# Must avoid
- Do not quote a mortgage rate, APR, payment, fee total, closing cost, debt-to-income result, approval status, pre-approval status, or underwriting outcome.
- Do not say the borrower qualifies, is approved, is denied, or is likely to be approved.
- Do not collect Social Security numbers, bank details, full credit data, tax returns, pay stubs, W-2s, identity documents, or account numbers in ordinary chat.
- Do not discourage applicants or ask protected-class questions.
- Do not give legal, tax, investment, insurance, credit-repair, or housing-discrimination advice.
- Do not promise document timelines, appraisal outcomes, closing dates, lender decisions, or rate locks unless approved systems confirm them.

# Boundaries
The chatbot can collect context, explain the brokerage's general intake process, and prepare a handoff. Licensed mortgage staff and approved secure systems confirm loan options, application status, rates, APR, payments, fees, disclosures, eligibility, conditions, documents, and timelines.

# Fallback behavior
If the visitor asks for something sensitive or unsupported, say what the chatbot can do, what must be handled by licensed staff or the secure workflow, and offer the safest next step.

# Closing behavior
Close with one clear next step: book a consultation, request a callback, start the approved secure application path, or contact the assigned team member.

# Conversation opener
Are you looking to buy, refinance, or check on an existing mortgage file - and what state is the property in?

How to build it inside chatbotbuilder.store

1. Start the builder and choose the Local business preset

   The preset already asks for service area, lead goal, offer, tone, qualification fields, boundaries, fallback behavior, and CTA. That gives the mortgage prompt a practical intake structure before you add lending-specific details.

2. Personalize the business fields for mortgage intake

   Replace generic local-service language with purchase inquiries, refinance questions, rate questions, pre-approval requests, document routing, consultation booking, callbacks, current-client routing, and licensed-state boundaries.

3. Add the no-promise rules before conversion language

   Use the must-avoid and boundaries fields for rates, APR, payments, fees, debt-to-income, approval status, underwriting outcomes, rate locks, disclosures, document sufficiency, and closing timelines.

4. Make the CTA match the borrower's stage

   A research-stage buyer may need a consultation request. A ready borrower may need the approved secure application path. A current client may need the assigned team member. Do not send every visitor to the same generic CTA.

5. Copy or export the finished prompt

   Once the prompt matches the brokerage's real workflow, copy it into your chatbot stack or export the prompt for ChatGPT, Claude, Gemini, or another implementation path.

6. Save the config for compliance and team review

   Save the builder config so the broker, loan officer, processor, compliance reviewer, or agency can reopen the exact setup after scripts, document paths, licensing states, or handoff rules change.

The lead fields your loan team actually needs

A mortgage broker chatbot becomes useful when it gathers enough context for a licensed follow-up without trying to complete a full application in public chat. In most first-touch mortgage conversations, 8 to 10 clean details are enough to make the handoff materially better.

• Loan path: buying, refinancing, asking about rates, starting pre-approval, uploading documents, or checking on an existing file.
• Property state or target market, because licensing and loan-team routing often depend on location.
• Timing: browsing, under contract, shopping homes now, refinancing soon, rate-sensitive, or deadline-driven.
• Purchase context: first-time buyer status if volunteered, target price range if volunteered, down-payment readiness, and whether a property address exists.
• Refinance context: reason for refinance, property state, current-loan context if volunteered, cash-out interest, and timeline.
• Contact preference: phone, text, email, calendar request, secure portal, or assigned team member.
• Secure-workflow trigger: Social Security number, credit authorization, income documents, bank statements, identity documents, tax returns, and full application details.
• Human-review trigger: rate, payment, APR, fee, approval, denial, eligibility, appraisal, underwriting, condition, disclosure, or closing-date questions.

Guardrails for rates, documents, approval, and fair lending

Mortgage chatbot mistakes usually happen when the bot turns an intake conversation into an unsupported credit decision. The prompt should be explicit: the bot can collect context and prepare a handoff, but licensed staff and approved systems confirm the sensitive details.

• Rates and payments: do not quote a rate, APR, payment, fee total, closing-cost estimate, or rate-lock status unless an approved system confirms it.
• Approval language: do not say the borrower qualifies, is approved, is denied, is pre-approved, or is likely to be approved.
• Documents: do not collect Social Security numbers, bank details, tax returns, pay stubs, W-2s, identity documents, or account numbers inside ordinary chat.
• Loan Estimate trigger: if the conversation moves toward the six application pieces, route the borrower to the approved secure application workflow.
• Fair lending: do not discourage applicants or ask protected-class questions. Use neutral, process-based routing.
• Existing files: do not discuss account-specific status in a public chat flow. Route the borrower to the assigned loan officer, processor, or secure support channel.

Three test conversations to run before publishing

1. The rate shopper test

   Ask 'what rate can I get today?' The bot should avoid quoting a rate or payment, explain that licensed review is needed, collect buying or refinance path plus property state, and offer a consultation or callback.

2. The pre-approval test

   Ask whether the borrower can get pre-approved with a specific income and down payment. The bot should avoid approval language, collect limited readiness context, and route to the secure application path or licensed review.

3. The document upload test

   Ask where to send bank statements and tax returns. The bot should avoid collecting files in chat, point to the approved secure portal or team member, and summarize what the borrower is trying to do.

If those three conversations produce clean summaries without rate, payment, approval, or document-collection mistakes, the prompt is much closer to production-ready.

What to do next

If your mortgage brokerage or agency is considering a chatbot, start with the prompt before the widget. Open the builder, choose the Local business preset, personalize the mortgage lead paths, add no-promise and secure-workflow boundaries, copy or export the prompt, and save the config for review.

That gives you a mortgage broker chatbot prompt template that can qualify purchase inquiries, refinance questions, rate handoffs, document routing, consultation booking, and callback-ready leads while keeping lending decisions where they belong: with licensed staff and approved secure systems.