Get started

The listing gets calls when the agent cannot answer

Real estate phone traffic follows the market, not the office calendar. A yard-sign call arrives while the listing agent is holding an open house. A relocation buyer calls at 9 p.m. after browsing listings all evening. A web lead submits an inquiry, waits a few minutes, and moves to the next search result. A closing coordinator spends an afternoon confirming an inspection window with a buyer, a lender, and an inspector who are never free at the same time.

Every one of those calls has a clear next step—give the property details, book the showing, capture the lead, confirm the appointment—but each one interrupts work that also cannot wait. Brokerages solve this with floor-duty rotations, answering services, and heroic phone habits, and calls still land in voicemail during the exact hours buyers are most active.

ThunderPhone gives a brokerage, team, or individual agent an AI phone agent that answers immediately, identifies itself as an AI assistant, works from the listing content and rules the team approves, and hands the conversation to a licensed agent whenever the call requires one.

This page is about brokerage and agent workflows—listings, showings, buyers, sellers, and transactions. If you manage rental properties and field maintenance or resident calls, see the separate property-management page instead.

Calls a real-estate agent workflow can cover

A ThunderPhone agent for a real estate team can be configured to:

  • answer sign calls and listing inquiries with details drawn from approved listing sheets in an attached knowledge base—price, beds and baths, square footage, HOA and disclosure basics, showing instructions;
  • schedule and confirm showings through a connected calendar or showing-management API, including buyer-agent showing requests on the team's listings;
  • capture buyer and seller lead intake—name, contact details, timeline, financing status, property of interest, how they found the number—and route the lead by rules the team defines, such as listing agent, buyer's-agent rotation, or geographic farm;
  • call web leads back within minutes of a form submission, triggered by the team's site or CRM, so a human conversation starts while the inquiry is still fresh;
  • place outbound open-house follow-up calls to visitors who signed in and agreed to be contacted, using the team's own script;
  • make transaction-milestone reminder calls from the team's own transaction data—inspection windows, appraisal appointments, contingency deadlines, final walkthrough, closing date and documents to bring;
  • coordinate with vendors the team already works with, confirming times with inspectors, photographers, stagers, and title or escrow contacts;
  • answer after hours and on weekends with the same intake discipline as midday;
  • speak with buyers and sellers in their preferred language, switching mid-call when a caller changes languages; and
  • run sphere-of-influence and past-client outreach campaigns under team-defined scripts and the customer's configured consent and compliance rules.

Every outbound workflow above operates on the team's own contacts and data—sign-in sheets, CRM records, transaction files—never on purchased or scraped calling lists.

Sample sign-call flow

Agent: Thank you for calling the Meridian Group. I am the team's AI assistant. Are you calling about a property, a showing, or something else?

Caller: I'm parked in front of your sign at 27 Alder Court. What's the price on this one?

Agent: 27 Alder Court is listed at $612,000. It is a three-bedroom, two-bath single-family home, about 1,850 square feet, with a two-car garage. Would you like to hear more, or set up a showing?

Caller: Is it a quiet street? What are the neighbors like?

Agent: That is a question I leave to a licensed agent rather than guess at. I can share what is in the listing, and I can have the listing agent call you back to talk through the area. Meanwhile—would a showing this week work?

Caller: Saturday morning, if that's possible.

Agent: Saturday at 10 a.m. is available. May I have your name and the best number for confirmations?

Caller: Dana, 555-0142.

Agent: You are set for Saturday at 10 a.m. at 27 Alder Court, and the listing agent will call you before then about your questions on the area. Anything else I can help with?

The agent answered from approved listing content, booked the showing, captured the lead—and deliberately declined to characterize the neighborhood, routing that question to a licensed agent instead.

Features that matter to a real estate team

A knowledge base fed by listing sheets

Upload listing sheets, disclosures summaries, buyer and seller FAQs, and office information as text, Markdown, CSV, PDF, or DOCX files (up to 5 MB per text file and 50 MB per PDF or DOCX), or import pages by URL. Documents are scoped per agent, so a single-listing line, a team line, and a brokerage main line can each search only their own approved material. When a listing changes, update the document once instead of re-briefing an answering service.

Lead routing the team controls

Routing is a set of rules the team writes: which questions qualify a lead, which answers trigger a transfer, and who receives which lead type. Cold transfer connects the caller immediately. On eligible production VoIP or SIP numbers, warm transfer puts the caller on hold while the AI agent privately briefs the human agent—name, property, timeline, financing status—and connects the two only after acceptance. If the agent declines or misses the transfer, the AI returns to the caller and follows the fallback the team configured.

Speed-to-lead callbacks from your own systems

An outbound call can be triggered through the API when a web form or CRM event fires, so a new inquiry gets a live conversation within minutes instead of a next-morning voicemail. HMAC-signed webhooks send the call outcome, transcript, and structured intake fields back to the team's systems, and MCP or API connections let the agent read and write CRM records mid-call where those systems expose the operations.

Campaigns with voicemail detection and retries

Open-house follow-up, milestone reminders, and sphere outreach run as campaigns over the team's contact lists inside configured calling windows. Voicemail detection can let the prompt decide, hang up, or leave a configured message, and campaigns can retry selected outcomes on a schedule. When a call reaches a lender's or title office's phone tree, the agent can navigate keypad menus and wait on hold.

47 languages with mid-call switching

The current catalog covers 47 languages and 123 voices. An agent starts in its primary language and can switch automatically when a caller uses a configured additional language, so a team serving multilingual markets can run one number instead of one per language. A web widget can also put the same agent on the team's listing pages for browser-based voice conversations.

Rehearse before the phone rings

Simulated AI callers can run graded scenarios before launch: the sign call, the lowball question, the caller who asks about schools and safety, the past client, the wrong number. Regression suites with pass-rate gates catch drift when prompts or listings change, and A/B experiments compare greeting or routing variants on live traffic. Calls produce transcripts, optional recordings, and grading for review.

What would this cost each month?

Suppose a brokerage handles 1,200 AI-assisted calls per month, averaging 3 minutes each:

1,200 calls × 3 minutes = 3,600 minutes per month

  • Spark: 3,600 × $0.02 = $72/month
  • Bolt: 3,600 × $0.05 = $180/month
  • Storm: 3,600 × $0.09 = $324/month

That is base engine usage, not a quote. Premium voice or language paths can add up to 3¢ per minute, and phone-number or carrier charges depend on the telephony setup. Hold time bills a flat 2¢ per minute; voicemail and phone-menu minutes bill at the engine rate. Optional add-ons such as verbal acknowledgements, call supervision, and long prompts carry their own published surcharges. The agent builder shows the current all-in rate before calls are placed.

The arithmetic is illustrative. Actual volume depends on listing count, lead sources, and how much outbound work the team routes through campaigns.

Guardrails for a regulated industry

Real estate calling sits inside two serious bodies of regulation, and the honest framing is that ThunderPhone provides the controls while the brokerage provides the compliance decisions.

Outbound calling and texting are heavily regulated. Do-not-call registries, prior-consent requirements, calling-hour restrictions, and state telemarketing laws all apply to real estate outreach, and regulators treat AI-generated voices as automated calls. On ThunderPhone, outbound campaigns run only under the customer's configured consent and compliance rules: the team defines who may be called, on what consent basis, in which time windows, and with what identification and opt-out handling. Before covered organizations enable outbound calling, an organization admin must complete ThunderPhone's one-time compliance confirmation. Configuring those rules correctly—and maintaining consent records and do-not-call suppression—is the customer's responsibility. ThunderPhone is not a source of scraped or purchased lead lists, and this page does not describe cold-calling them.

Fair-housing rules shape what the agent may say. Steering-sensitive topics—neighborhood character, demographics, schools, safety—are questions a licensed professional should handle under the brokerage's fair-housing policy. A ThunderPhone agent can be instructed to answer property questions only from approved listing content, to give no opinions about neighborhoods or the people who live in them, and to route those questions to a licensed agent, as the sample call shows. The brokerage authors that policy and should test it in simulation before launch; ThunderPhone supplies the enforcement surface, not legal advice.

ThunderPhone's platform posture is GDPR and HIPAA compliant, and customers remain responsible for configuring their workflows, recordings, data handling, and notices for their own obligations, including state call-recording consent rules.

Frequently asked questions

Can an AI phone agent answer questions about my listings?

Yes, from content you approve. Attach listing sheets and FAQs to the agent's knowledge base and it can search them mid-call to answer price, feature, and showing-logistics questions. It answers from those documents rather than general knowledge, and the team should update documents when listing details change.

Can it schedule showings?

Yes, when connected to an approved calendar or scheduling API. The agent can offer available windows, book the showing, capture the caller's contact details, and repeat the confirmation. Campaigns or individual outbound calls can then confirm or reschedule as the date approaches.

How does lead routing work for a team?

The team defines the rules. Intake questions qualify the caller, and the configuration decides who gets the call or the lead record—listing agent for sign calls, a rotation for new buyers, a named agent for a geographic farm. Warm transfer on eligible production numbers lets the AI brief the human agent privately before connecting the caller.

Is outbound calling to leads and past clients legal?

Outbound real estate calling is regulated by do-not-call registries, consent requirements, calling-hour limits, and state telemarketing laws, and automated or AI-voice calls generally require prior consent. ThunderPhone requires a one-time admin compliance confirmation before covered organizations enable outbound calling, and provides configurable calling windows, identification, and campaign controls. The customer remains responsible for consent, list hygiene, do-not-call suppression, and every other applicable rule; consult your broker or counsel before launching a campaign.

Will the agent answer questions about neighborhoods or demographics?

It should not, and it can be configured not to. Fair-housing rules make neighborhood characterizations and demographic commentary steering-sensitive, so the recommended configuration limits the agent to approved listing content and routes those questions to a licensed agent. Test that boundary in simulation before going live.

Can it call buyers with transaction reminders?

Yes, using the team's own transaction data. Milestone calls—inspection scheduled, appraisal window, contingency deadline, final walkthrough, closing time and location—can run as individual outbound calls or campaigns to parties who have consented to be called, with voicemail handling and retries configured by the team.

Can it speak Spanish or other languages with my buyers?

The current catalog supports 47 languages, and an agent can switch automatically mid-call when a caller uses a configured additional language. Premium voice or language paths can add up to 3¢ per minute, and the selected voice must support the configured language set.

What happens after hours?

A deployed agent answers calls routed to its number at any hour with the same scripts, knowledge, and routing rules as daytime. It can book showings within allowed availability, capture complete lead intake, and flag time-sensitive calls for the on-duty agent according to the team's escalation policy.