What a real estate AI agent actually does
Reads a new lead the moment it lands, qualifies it by timeline, budget and area, and either books a showing or hands it to a human agent with full context.
A real estate AI agent sits in front of a brokerage's lead sources, a Zillow Premier Agent lead, a portal inquiry or a website form, and runs a loop: capture the lead, text or call them within seconds, ask qualifying questions on timeline, budget, pre-approval and area, check the MLS/IDX feed through the RESO Web API for listing data, and either book a showing on the agent's calendar or hand a qualified lead to a human with notes attached. The agent lives inside or alongside a CRM like Follow Up Boss, kvCORE/BoldTrail or Sierra Interactive, and an AI ISA such as Structurely layers speed-to-lead texting and calling on top of whichever CRM already holds the lead.
Speed to lead is the single most-cited number in this category. The 2007 Lead Response Management study found that contacting a lead within 5 minutes instead of 30 makes it 100 times more likely to be reached and 21 times more likely to be qualified, and a 2011 Harvard Business Review analysis of 2,241 companies found that responding within an hour made firms roughly 7 times more likely to qualify a lead than waiting longer. An AI agent that texts or calls the instant a lead hits the CRM is built specifically to chase that gap, which is why response-time automation, not just chat, is the core pitch of every vendor in this space.
Where agents still fail: judgment about a specific listing's condition or a buyer's financing nuance, multi-offer negotiation strategy, and any conversation that drifts into steering, commenting on a neighborhood's safety, schools or demographics, which the Fair Housing Act treats as a discrimination risk regardless of whether a human or an AI said it. An agent nurturing a database of thousands of past leads can also quietly send the same generic message to everyone, which erodes trust faster than it builds pipeline.
AI capabilities involved
Instant lead response by text or voice
Multi-turn qualification on budget, timeline and area
Showing scheduling against a calendar and MLS listing data
Long-term nurture and database reactivation messaging
Who uses this
- Solo agents and small teams getting more inbound leads than they can call back within minutes
- Brokerages running a dedicated ISA function that wants to extend coverage after hours
- Teams on Follow Up Boss, kvCORE/BoldTrail or Sierra Interactive deciding whether to add AI qualification on top
- New-construction and high-volume listing teams needing consistent follow-up across a large, aging lead database
What the real estate agent handles (and what it shouldn't)
| Task | Agent handles it? | Systems it touches | Human role |
|---|---|---|---|
| New portal lead response (Zillow, a website form) | Fully | CRM, MLS/IDX feed | None unless the lead asks a complex financing question |
| Lead qualification (timeline, budget, pre-approval) | Fully | CRM, Follow Up Boss or kvCORE | Reviews before a showing is booked |
| Showing scheduling | With approval | Calendar, ShowingTime | Confirms the final time with the client |
| Database reactivation and nurture texts | Fully | CRM | Pulls any flagged VIP or past-client leads out of automation |
| Listing Q&A (price, features, HOA) | Fully | MLS/IDX feed | None unless the question needs seller input |
| Open-house follow-up | Fully | CRM, calendar | None |
| Financing and offer-strategy conversations | No | CRM | Takes over immediately |
| Neighborhood, school or safety questions | No | CRM | Redirects to public resources only; never comments to avoid steering |
Off-the-shelf AI agents on the market
Real products you can sign up for today — with current 2026 pricing, honest pros and cons.
Follow Up Boss
Teams that want lead qualification built into the CRM they already use
No free tier published; a live demo is available
$69/user/mo (Grow); $499/mo for 10 users (Pro); $1,000/mo for 30 users (Platform)
Platform tier scales to 30 users at $1,000/mo
Pros
- +AI qualification features included on every plan tier
- +Deep CRM feature set most real estate teams already use
- +Scales from a single agent to a 30-person team
Cons
- −Pricing is reported by third parties, not an official vendor price page, so confirm current tiers before buying
- −AI features sit inside a full CRM, so you're paying for more than just the agent if that's all you want
Structurely (AI ISA)
Teams that want a dedicated AI ISA layered on top of an existing CRM
No public free tier
Reported around $179/mo on its base plan, or $499/mo plus roughly $0.12 per action and a $2,500 setup fee on its higher tier (sources conflict; get a written quote)
Custom for high-volume teams
Pros
- +Purpose-built AI ISA for real estate speed-to-lead texting and calling
- +Works alongside whichever CRM already holds the lead
- +Moved to a credit-based model in April 2026, which can lower cost for lighter lead volume
Cons
- −Public pricing is inconsistent across third-party sources
- −Per-action billing is harder to forecast than a flat monthly fee
Ylopo
Teams that want lead generation and AI nurture from one vendor
No public free tier; a demo is required
Reported around $295/mo for its AI product alone; $1,000+/mo once ad spend is included
Custom for brokerage-wide rollouts
Pros
- +Combines AI lead nurture with its own paid-ad lead generation
- +Built specifically for real estate's listing and nurture workflows
- +Established vendor with a large brokerage client base
Cons
- −Public pricing is a third-party estimate, not an official number
- −Its value is tied to also buying Ylopo's ad spend, which raises the real monthly cost
The AI stack
A real estate agent needs a fast reasoning model for instant response, access to the MLS/IDX feed and CRM, real tool access to book a showing, and a way to catch a fair-housing or TCPA mistake before it reaches a lead.
Reasoning model
Drives the instant first response and the multi-turn qualification conversation that follows
Claude Haiku 5.5
$0.10/M input tokens (up to 100K), $0.50/M outputThe first-touch response
Claude Sonnet 5.5
$2/M input, $10/M outputMulti-turn qualification once a lead replies
GPT-6.1 Sol
$2/M input, $10/M outputTeams already standardized on OpenAI
Our pick: Claude Haiku 5.5 or GPT-6 Luna for the first-touch text or call, with Sonnet 5.5 or GPT-6.1 Sol handling multi-turn qualification once a lead responds.
Knowledge & retrieval
Grounds listing answers in real MLS data and qualification scripts in your brokerage's written language
MLS/IDX feed via the RESO Web API (native)
Included in your existing MLS/IDX accessListing Q&A
CRM-native nurture templates
Included in your CRM planQualification and nurture
Our pick: Pull listing data straight from your MLS/IDX feed through the RESO Web API, and keep qualification scripts in your CRM so a human can audit exactly what the agent is saying.
Integrations / tools
Lets the agent actually text, call, book a showing and log back to the CRM
CRM-native connectors (Follow Up Boss, kvCORE)
Included on most tiersBuy-SaaS
Calendar/ShowingTime plus SMS and voice APIs for custom builds
Developer time plus carrier feesCustom builds and n8n
Our pick: Use your CRM's native AI ISA integration until you need a showing-scheduling or multi-brokerage workflow it doesn't support.
Observability & evals
Catches a fair-housing or TCPA mistake before it reaches more leads
Vendor dashboards
IncludedBuy-SaaS
Langfuse plus a human fair-housing script review
Free up to 50K units/mo, then $29+/moCustom builds
Our pick: Add a human fair-housing review of your qualification scripts before launch, then log every conversation and sample it weekly.
Reference architecture
The hardest engineering problem is staying inside the 5-minute response window while still asking fair-housing-safe questions: the agent has to qualify on timeline, budget and area without ever drifting into commentary on neighborhoods, schools or safety, which the Fair Housing Act treats as steering regardless of intent.
Lead arrives
Portal, website form or CRM webhookA new lead from Zillow Premier Agent, a website form or a paid campaign lands in the CRM
Instant response
Fast reasoning modelA text or call goes out within seconds, before the 5-minute window that drives the 100x contact-rate gap closes
Qualification
Reasoning model plus CRMThe agent asks about timeline, budget, pre-approval and area, logging answers to the CRM
Listing lookup
MLS/IDX feedRelevant listings are pulled through the RESO Web API to answer specific questions
Showing booking or handoff
Calendar, ShowingTimeA qualified lead books a showing directly, or gets handed to a human agent with full notes
Nurture
CRM automationUnqualified or not-yet-ready leads enter a longer-term nurture sequence instead of being dropped
Logging and eval
Observability toolEvery conversation is logged and a sample is reviewed weekly for fair-housing and TCPA compliance
Estimated cost per request
~$0.03-$0.08 per qualified lead conversation in model costs alone, before any SMS or voice carrier fee
Cost calculator
Drag the sliders to model your actual usage. The numbers update in real time so you can stress-test economics before writing a single line of code.
Estimate your monthly cost at your own lead volume and outbound contact rate.
Estimated monthly cost
$202
≈ $2,418 per year
Calculator notes
- The per-action rate assumes Structurely's reported roughly $0.12/action pricing at an 80% outbound-contact rate; move the contact-rate slider to see how your own mix changes the real action count.
- Model cost covers the reasoning calls behind instant response and qualification, and applies whether you buy SaaS or build custom.
- A custom build swaps the platform fee and per-action rate for your own model, SMS and voice infra bill, typically $150-$600/mo at this volume (see build_vs_buy).
- TCPA consent capture and DNC scrubbing add legal and process overhead on every path, not just DIY.
Build it yourself with vibe-coding tools
n8n gets a working speed-to-lead agent running in an afternoon: a webhook on new CRM leads, an AI Agent node with SMS and calendar tools, and a qualification script that logs back to the CRM.
Time to MVP
1-3 days for a first version, 2-4 weeks to production quality with TCPA consent handling and fair-housing review
Total cost to MVP
$0-$70 (n8n Cloud Starter plus model and SMS API costs)
You'll need
Starter prompt
Set up a webhook trigger on new leads from your CRM. Feed the lead into an AI Agent node using Claude Haiku 5.5 for the instant first text, escalating to Claude Sonnet 5.5 for multi-turn qualification once the lead replies, with a Twilio tool for SMS and an HTTP Request tool for your MLS/IDX feed through the RESO Web API. Give the AI Agent node this system prompt: 'You are a real estate ISA for {Brokerage Name}. Text the lead within seconds of assignment, confirm you are an AI assistant, and ask about their timeline, budget range and preferred area one question at a time. Never comment on a neighborhood's schools, safety or demographics, and never discuss financing specifics or offer strategy: if asked, say a licensed agent will follow up. Use the MLS tool to answer specific listing questions. Once timeline, budget and area are confirmed, offer to book a showing on the calendar tool. If the lead goes quiet for more than a day, hand off to the long-term nurture sequence instead of repeating yourself.' Add a Switch node that routes a booked showing to the assigned agent's calendar and logs every qualifying answer back to the CRM record.
Paste this into n8n
Follow-up prompts (run in order)
- 1
Add a daily digest of leads qualified but not yet booked
- 2
Add a DNC-scrub check before every outbound text or call
- 3
Add a long-term nurture sequence for leads who aren't ready yet
- 4
Add a fallback to a human ISA if the lead doesn't respond to two automated attempts
Expected output
A working agent that texts a new lead within seconds, qualifies them against your script, and either books a showing or hands them to a human with full notes attached.
Known gotchas
- !TCPA consent has to be captured and logged before the first outbound text or call, not after
- !A generic qualification script that never varies starts to feel robotic to leads who talk to more than one agent at your brokerage
- !Webhook retries from some CRMs can trigger duplicate outbound texts, so dedupe on the lead ID
- !There's no built-in fair-housing review: you have to manually audit scripts for steering language yourself
- !MLS/IDX feed access and formatting vary by region, so test listing lookups against your specific feed before launch
Compliance & risk reality check
A real estate agent that qualifies and contacts leads carries real regulatory weight on two fronts: fair housing and outbound consent.
Fair Housing Act steering
The agent must never comment on a neighborhood's schools, safety, crime or demographics, or give inconsistent answers to different prospects, both of which the Fair Housing Act treats as discriminatory steering regardless of whether a human or an AI said it.
Mitigation: Write and fair-housing-review the qualification script before launch, give every prospect the same answers, and log every conversation for audit.
TCPA consent for outbound AI calls and texts
Outbound AI voice and text to leads needs prior express consent under the FCC's 2024 ruling that AI-generated voices count as artificial voices under the TCPA, with penalties up to $1,500 per willful violation.
Mitigation: Capture written consent naming AI and automated calls at lead capture, scrub against DNC lists, and disclose the AI at the start of every call.
AI disclosure
A lead who asks whether they're talking to a bot should get a truthful answer, and several states now require disclosure outright.
Mitigation: State "I'm an AI assistant" in the first message or call, which also satisfies California, Utah and Maine's disclosure laws.
Vendor training on your lead data
AI ISA vendors process your lead conversations to generate responses, and some contracts allow using that data to improve their own models.
Mitigation: Add a no-training clause to vendor contracts where offered.
Build vs buy: the real math
3-6 weeks for a RapidDev build; 1-3 days for a basic n8n version
Custom build time
$13K-$25K for an agency build (up to $35K for multi-brokerage work); $0-$70/mo for DIY
One-time investment
Roughly 400-500 new leads a month, where a per-action AI ISA fee exceeds the amortized first-year cost of a custom build
Breakeven vs buying
At 150 leads a month and an 80% contact rate, a Structurely-style setup at roughly $179/mo base plus $0.12 per action runs about $193/mo, dramatically cheaper than a $13K+ agency build in year one. At 500 leads a month the per-action fees alone approach $480/mo on top of the base platform fee, and a custom agent's marginal cost per extra lead, around $0.03-$0.08 in model tokens, starts to look competitive once a team also needs multi-brokerage routing or a non-standard CRM integration a packaged ISA vendor doesn't support.
Skip the DIY — RapidDev builds the production version
A Lovable MVP gets you a demo. Production needs auth that doesn't leak data, AI calls that don't bankrupt you, observability when models drift, and code you can audit. That's what we ship through our AI app development service.
Discovery call (free)
30 minWe map your exact real estate AI agent use case: who uses it, target volume, AI model choice, integrations, compliance scope. You get a detailed scope document and fixed-price quote within 48 hours.
AI-accelerated build
3-6 weeks for a RapidDev build; 1-3 days for a basic n8n versionOur engineers use Claude Code, Lovable, and custom tooling to ship 3–5x faster than agencies. You see weekly progress in a staging environment — not a black box.
Launch + handoff
1 weekWe deploy to your infrastructure, transfer the GitHub repo, set up CI/CD and monitoring, and train your team. You own 100% of the source code, prompts, and model configurations.
What you get
Timeline
3-6 weeks for a RapidDev build; 1-3 days for a basic n8n version
Investment
$13K-$25K for an agency build (up to $35K for multi-brokerage work); $0-$70/mo for DIY
vs SaaS
ROI in Roughly 400-500 new leads a month, where a per-action AI ISA fee exceeds the amortized first-year cost of a custom build
30-min call. Fixed-price quote within 48 hours. No commitment.
Frequently asked questions
How much does an AI agent for real estate cost?
Follow Up Boss runs $69-$1,000/mo by team size, Structurely is reported around $179-$499/mo plus roughly $0.12 per action, and Ylopo is reported around $295/mo and up with ad spend. A custom RapidDev build runs $13K-$25K upfront (up to $35K for multi-brokerage work), plus $150-$600/mo in model and infra costs. A DIY n8n version costs $0-$70/mo.
How long does it take to set up?
A SaaS AI ISA connected to your existing CRM takes 1-2 weeks. A custom agency build takes 3-6 weeks because it has to be wired into your CRM, MLS feed and written scripts. A basic n8n version can work within 1-3 days, though TCPA consent handling and fair-housing review take 2-4 weeks to get right.
Can RapidDev build this for us?
Yes. RapidDev has shipped 600+ apps and offers a free 30-minute consultation to scope your lead sources, CRM and compliance needs before you commit to anything.
Will leads know they're talking to AI?
They should, and several states require it outright. State "I'm an AI assistant" in the first text or call, which satisfies disclosure rules in California, Utah and Maine and builds trust faster than letting a lead figure it out themselves.
Can an AI agent actually book a showing?
Yes, once it has confirmed timeline, budget and area, a well-built agent can book directly on a connected calendar or through ShowingTime, then hand the booked showing to the assigned agent with full qualification notes.
What questions should the agent never ask or answer?
Anything about a neighborhood's schools, safety or demographics, and anything involving financing specifics or offer strategy. Those are Fair Housing Act steering risks and judgment calls that belong with a licensed human agent, not an AI script.
How fast does the agent actually need to respond?
As close to instant as possible. The 2007 Lead Response Management study found contacting a lead within 5 minutes instead of 30 makes it 100 times more likely to be reached and 21 times more likely to be qualified, which is the entire reason this category exists.
Can it work with Follow Up Boss, kvCORE or Sierra Interactive?
Yes. Follow Up Boss has AI qualification features on every plan, and AI ISA vendors like Structurely layer on top of whichever CRM already holds the lead. A custom build can connect to any of these plus systems no SaaS vendor supports.
Want the production version?
- Delivered in 3-6 weeks for a RapidDev build; 1-3 days for a basic n8n version
- You own 100% of the code
- AI cost monitoring built in
30-min call. No commitment.
