Help users explore mortgage options by collecting key inputs like purchase price, credit score, down payment, and property location, then presenting tailored loan choices with current rates.
Use the channels below to try the Home Mortgage Agent live. Chat with it in your browser, call it by phone, or send an SMS.
On your phone? Tap to call and speak directly with the mortgage agent. Try asking about 30-year fixed rates, FHA loan options, or what you'd qualify for with a 720 credit score.
Prefer texting? Send an SMS to start a conversation. Ask about down payment requirements, compare loan types, or get a quick rate estimate.
See exactly how this agent was built - the prompt design, tool integrations, and configuration that power the mortgage experience.
Watch a step-by-step walkthrough of building this agent from scratch on the Syllable platform.
The Home Mortgage Agent helps potential home buyers understand their financing options by collecting key inputs - purchase price, credit score, down payment, and property location - and presenting tailored loan choices with current interest rates.
The agent integrates with Bankrate.com to pull up-to-date rate information, then walks the caller through relevant loan types such as 30-year fixed, 15-year fixed, FHA, and adjustable-rate options. It explains how each factor affects monthly payments and qualification, keeping the conversation approachable without oversimplifying.
This agent demonstrates how the Syllable Agentic Platform connects AI agents to external financial data through tool integrations and declarative prompt configuration. The same agent handles voice calls, SMS, and chat with no separate setup per channel.
The Home Mortgage Agent is not just a chatbot that retrieves rates. Its prompt embeds real financial compliance logic that shapes every turn of the conversation.
The agent follows a strict 7-step sequence: ZIP code, purchase versus refinance, property price, down payment, occupancy type, credit score band, and timeline. It never combines questions. Each turn asks one thing and waits. This conversation design pattern ensures clean, unambiguous data capture in regulated domains.
The agent applies real financial math: adds 0.5% in lender fees for APR, classifies loans as JUMBO when they exceed the local high-balance limit (adding 0.25%), and adjusts rates by 0.40% for credit scores below 700. It screens across Conventional, FHA, VA, USDA, and Jumbo programs, then ranks the top three by lowest payment, lowest cash-to-close, and lowest total interest.
The prompt explicitly prohibits requesting a full Social Security number, making rate predictions, steering to specific lenders, or storing income and debt details. Every rate quote includes a regulatory disclaimer. These rules live in the prompt itself with no backend code enforcing them.
Structured data collection, financial calculations, and compliance guardrails apply across real estate, lending, and property management.
Answers tenant questions about units, pricing, pet policies, and move-in costs. Collects application information and schedules tours. A knowledge base holds property-specific details the same way the Shopify Agent uses a store FAQ data source.
Collects property details and queries multiple rate APIs to present side-by-side coverage options. Same structured single-question data collection pattern as the mortgage agent, adapted for insurance qualification.
An outbound agent that contacts all parties in a real estate closing - title company, lender, inspector, appraiser - to check on document status and deadlines. Uses the same outbound campaign pattern applied to transaction management.
The agent uses a platform tool that integrates with Bankrate.com to retrieve up-to-date interest rate information. Rates reflect current market conditions at the time of your conversation.
Yes. The agent explains the differences between common mortgage products - fixed-rate, adjustable-rate, FHA, and conventional loans - and shows how each option affects monthly payments based on the inputs you provide.
This is a demo agent designed to illustrate agent capabilities. It does not store personal financial information or make actual lending decisions. In production, the Syllable Agentic Platform enforces data encryption, access controls, and audit logging for every interaction.
Yes. The documentation walks through every step of building this agent, from prompt design and tool integration to deployment across voice, SMS, and chat.
The agent screens your inputs against five loan programs (Conventional, FHA, VA, USDA, Jumbo), applies credit-score adjustments and lender fees, then ranks the top three options. It presents them by lowest monthly payment, lowest cash-to-close, and lowest total interest so you can compare trade-offs at a glance.
The agent never asks for your full Social Security number, income details, or debt amounts. It works with credit score bands rather than exact scores. These restrictions are enforced through the prompt itself, demonstrating how compliance rules can be embedded directly in agent configuration.