Skip to main content

Real Estate Agent

In this tutorial, you’ll build a voice agent for a real estate agency. It handles inbound calls from buyers and renters, learns their preferences, matches them to properties in your knowledge base, schedules showings, and warm-transfers serious buyers to a human agent.

What You’ll Build

A phone agent that:
  • Greets callers and identifies whether they’re buying, renting, or selling
  • Asks about preferences: location, budget, bedrooms, must-haves
  • Searches your property listings knowledge base for matches
  • Describes properties conversationally (not like a database dump)
  • Schedules property showings on Google Calendar
  • Sends property links and showing confirmations via SMS
  • Warm-transfers serious, qualified buyers to a human agent
  • Handles after-hours calls and captures leads 24/7
Time to set up: ~10 minutes

Step 1: Create the Agent

  1. Go to your thinnestAI Dashboard
  2. Click Create Agent
  3. Name it: Property Inquiry Agent
  4. Set the description: Handles property inquiries, matches listings, and schedules showings

Step 2: Choose the Model

Select Claude Sonnet as your model. Claude Sonnet excels at conversational property descriptions — it can take dry listing data and present it naturally and engagingly over the phone. It also handles the nuanced multi-step flow of qualifying buyers and scheduling showings.

Step 3: Add Knowledge

Property Listings

Upload your active property listings. Format them clearly so the agent can search and match effectively:
Upload as many listings as you have. The agent will search across all of them.

Agent Availability

Upload a document with your team’s availability and specialties:

Neighborhood Information

Upload a guide with neighborhood details callers commonly ask about:
Upload all three documents in the Knowledge tab.

Step 4: Assign Tools

Enabled automatically with your uploaded listings and neighborhood data.

Google Calendar

For scheduling property showings.
  1. Go to the Tools tab
  2. Enable Google Calendar
  3. Connect the calendar(s) for your agents (Sarah, Marcus, Lisa)
  4. The agent will check availability and create showing appointments

SMS

For sending property links and showing confirmations.
  1. Enable SMS in the Tools tab
  2. Configure your agency phone number

Step 5: Set the System Prompt

Step 6: Assign a Phone Number

  1. Go to the Phone tab
  2. Assign a phone number (use your agency’s main line or a dedicated inquiry number)
  3. Set call direction to Inbound
  4. Save

Step 7: Test Your Agent

Testing Checklist

What to Listen For

  • Property descriptions: Does the agent paint a picture, or just list specs?
  • Natural flow: Does preference gathering feel conversational or like a questionnaire?
  • Market knowledge: Does the agent share useful neighborhood context?
  • Enthusiasm: Does it sound genuinely helpful and interested?
  • Transfer timing: Does it recognize when a caller is serious enough to transfer?

Tips for Improvement

  1. Keep listings updated. Stale listings are worse than no listings. Update your knowledge base weekly or set up an integration with your MLS feed.
  2. Add photos via SMS. When texting a listing link, the property page should have great photos. The agent can say “I just texted you the listing — take a look at the photos and let me know what you think.”
  3. Track lead sources. Use different phone numbers for different marketing channels (Zillow, your website, yard signs). This tells you which channels drive the most qualified leads.
  4. Follow-up campaigns. Set up a campaign to call back leads who inquired but didn’t schedule a showing within 3 days. A gentle follow-up converts a surprising number of leads.
  5. Seasonal knowledge. Update your neighborhood descriptions seasonally. Mentioning the farmers market in summer or holiday lights in winter adds a personal touch.
  6. Open house reminders. Before an open house, run a campaign calling recent inquirers in that area: “Hi [Name], just a quick heads up — the property on Evergreen Terrace has an open house this Saturday from 1 to 3. Would you like to stop by?”
  7. Multi-language support. If your market has a significant non-English-speaking population, consider creating a second agent with a system prompt in that language.

These examples cover the most common voice agent use cases. Each one can be customized to fit your specific business. Start with the example closest to your needs, test it thoroughly, and iterate based on real call data. Need help with a use case not covered here? Reach out to our team at support@thinnest.ai.