Most open house follow-up is terrible. An agent sends one generic “Thanks for stopping by!” email, waits a week, and wonders why nothing converts. The visitors who were actually interested — the ones who asked about school districts and took photos of the kitchen — got the same email as the neighbors who came out of curiosity.
This playbook fixes that. You’ll get exact ChatGPT prompts for a four-touch follow-up sequence, a timing schedule that matches how buyers actually make decisions, and notes on plugging this into your CRM so it runs without you.
What you need before you start
- Sign-in sheet data (name, email, phone, and any notes you jotted during the showing)
- Your rough impression of each visitor: hot lead, casual interest, neighbor/nosy, or unclear
- A CRM that can send emails and texts on a schedule (Follow Up Boss, Lofty, kvCORE, or similar — see our Follow Up Boss vs. Lofty breakdown for comparison)
You will not be automating all of this with AI. The AI handles drafting; your CRM handles sending; you handle the actual human conversations that come back.
The 4-touch sequence and timing
| Touch | Channel | Timing | Goal |
|---|---|---|---|
| 1 | Same evening, 2–4 hours after open house closes | Warm acknowledgment, invite response | |
| 2 | Text | Next morning (8–9am) | Short, personal, nudge to schedule a showing |
| 3 | Day 3–4 | Deliver value (comparable recent sales, neighborhood info) | |
| 4 | Email or call | Day 10–14 | Long-tail check-in for the leads who went quiet |
The same-evening email is the most important one. Practitioner experience and industry benchmark data consistently show lead responsiveness drops sharply within the first hour of inquiry — and collapses further after 24 hours. Sending it from your phone on the drive home — or better, having your CRM fire it automatically — makes a real difference.
The prompts
Copy these exactly, then fill in the brackets. Run them in ChatGPT (free tier works fine for this).
Touch 1 — Same-evening email
Prompt:
Write a warm, personal follow-up email from a real estate agent to someone who attended an open house today at [address]. The visitor’s name is [First Name]. Notes from the showing: [paste your notes — e.g., “asked about school district twice, mentioned they’re renting nearby and lease is up in August”]. Keep it under 120 words. Do not use phrases like “it was so great to meet you” or “feel free to reach out.” End with one specific open question based on the notes. Tone: direct, warm, professional.
Why the negative instructions matter: ChatGPT defaults to filler phrases that sound like they came from a mass template. The explicit “do not use” list breaks that pattern. Add your own pet phrases to avoid.
Sample output (for a visitor who asked about schools):
Hi Sarah,
Thanks for coming through 412 Elm today. You asked about the school district twice — worth knowing that Lincoln Elementary is a 0.4-mile walk and the middle school just finished a renovation last year. That whole attendance zone has been steady for resale.
You mentioned your lease is up in August. What’s your timeline looking like — are you in active search mode or still exploring?
[Agent Name]
Touch 2 — Next-morning text
Prompt:
Write a follow-up text message (under 60 words) from a real estate agent to [First Name] who attended an open house yesterday at [address]. The agent wants to invite them to schedule a private showing. It should feel personal, not like a mass blast. Do not include a calendar link in the text itself — just invite them to reply to set a time. Notes about the visitor: [paste notes].
Texts should be short. A text that looks like an email gets ignored. If your CRM supports merge fields, the name and address can be dynamic.
Touch 3 — Day 3–4 value email
Prompt:
Write a follow-up email (under 200 words) from a real estate agent to [First Name] who attended an open house at [address] four days ago. The agent wants to provide useful market context without being pushy. Include: one recent comparable sale in the neighborhood (use this data: [paste a comp from your MLS — address, bed/bath, sold price, days on market]), and a note about current inventory in the area. Do not ask them to schedule a showing in this email. End with a soft door-open sentence. Tone: informative, low-pressure.
The comp data is the key ingredient. This touch works because it gives the prospect something real, not a sales pitch.
Touch 4 — Day 10–14 check-in
Prompt:
Write a short check-in email (under 100 words) from a real estate agent to [First Name], who attended an open house at [address] about two weeks ago and has not responded to previous follow-up. Acknowledge that timing might not be right, leave the door open without any pressure, and include one sentence of new information about the listing or market. Tone: completely relaxed, no urgency language. Do not include phrases like “just checking in” or “touching base.”
The two banned phrases appear in more than half of real estate follow-up emails. They signal low effort immediately.
CRM integration notes
Once you have the four templates drafted and lightly personalized, the goal is to get them loaded into your CRM as an action plan so you’re not manually sending each one.
In Follow Up Boss: Create a new Action Plan named “Open House Follow-Up.” Add each step with the channel (email/SMS), timing offset (0 days, 1 day, 3 days, 12 days from enrollment date), and paste the template text. Enroll each contact from the sign-in sheet manually after the open house — takes about 10 minutes in the car.
In Lofty/kvCORE: The workflow is similar. Create a campaign, set up the drip sequence, and trigger it from the contact record. Both platforms support SMS and email in the same campaign.
Hot leads vs. everyone else: Tag or segment your visitors before enrolling. Serious buyers (your “hot” tag) should get a direct call within the first 24 hours in addition to the email sequence, not instead of it. The email sequence is for the long tail.
One thing AI won’t do for you: handle the responses. When someone replies to Touch 1 or 2, that conversation is yours. The AI got them talking. You close.
For the full picture on building a CRM-powered follow-up system on a budget, see our solo agent AI stack under $200 post.
Common questions
How do I handle visitors who did not leave an email, only a phone number?
Run the same sequence with Text-only touches. Adjust Touch 3 to be a phone call instead of an email. The timing stays the same. Prompt ChatGPT for a call script outline instead of an email — the same prompt structure works with "write a 90-second call script" substituted for "write an email."
Can I use these prompts for broker open houses, not just public open houses?
Yes, with minor adjustments. Broker open follow-up should reference the specific agent or team by name ("Thanks for bringing your buyers through") and focus on cooperation and incentives rather than buyer motivation. Adjust the notes section of each prompt accordingly.
Will ChatGPT produce the same output every time?
No. AI outputs vary between runs even with identical prompts. Run each prompt 2–3 times and pick the best version, or combine the strongest parts of two outputs. The variation is actually useful — if you enroll 15 visitors, they will not all receive identical emails.
What if the visitor said something negative about the house?
Include it in the notes section of your prompt. For example: "Visitor mentioned the kitchen felt small." ChatGPT will factor it in and can frame the follow-up around value or alternative properties rather than pushing the same listing. That honesty builds trust faster than ignoring objections.
