Picture a buyer scrolling listings at 9:40 on a Tuesday night. They like a three-bed on Elm Street, fill in the enquiry form on two portals and your website, then go to sleep. By 8 the next morning one agent has already answered their question and offered two viewing slots on Saturday. In our experience, that’s usually the agent who gets the client.
Nobody on your team is slow on purpose. Enquiries arrive from portals, the website, Facebook ads, texts and missed calls, often while agents are driving between viewings. We’ve worked with property businesses for years, from the Retool apps Houst’s operations team runs its rental portfolio on to engineering work on MaverickRE’s sales analytics platform, and the fix is nearly always the same: get every lead into one place, answer it properly within minutes, and only then hand it to a person. To show what that looks like, we built the core of it as a workflow and ran two test enquiries through it.
The workflow at a glance
We built it in n8n, but nothing here depends on the tool. The same steps work in GoHighLevel workflows, HubSpot, Make or Zapier. For the test, a stand-in played the part of the CRM, the text-message provider and Slack, so no real messages went out.
Start by putting every enquiry in one place
Most follow-up problems begin in the inbox. Portal enquiries land in one mailbox, web forms in another, ad leads in a spreadsheet somebody exports on Mondays. Whatever isn’t in front of an agent gets forgotten.
So the first job is plumbing, not AI. Every source feeds one CRM the moment the lead arrives, with the source and the property attached. Along the way, clean up what each source sends: phone numbers in one format, the area worked out from the ZIP code, and a record of whether the person agreed to texts. It’s unglamorous work, and it’s where most of the gains come from.
Make the first reply useful
“Thanks for your enquiry, we’ll be in touch” is technically a reply, but it doesn’t move anything forward. A good first message names the property they asked about, offers two or three viewing times or a booking link, and is signed by a real agent. Here is the text our test run sent, about a second after the enquiry arrived:
Notice what it doesn’t say: “still available”. Unless the workflow checks the listing first, don’t let an automated message promise anything the agent might have to take back.
Route leads with a time limit
Decide who gets which lead before it arrives: by area, by listing, by buyer or seller, or taking turns where several agents cover the same patch. Then add the rule most teams skip. If nobody picks the lead up within a set time, say 15 minutes during working hours, it moves to the next agent and a manager gets a message.
That one rule does more for response times than any amount of reminding people in team meetings.
Qualify lightly
You don’t need a ten-question form. Two or three questions in the follow-up conversation are enough to sort leads: buying or renting, rough timeline, and whether financing is sorted or they still need to sell. Answers update the lead in the CRM automatically, so in the morning an agent sees who to call first instead of a long list in date order.
Let people book viewings themselves
Every “does Thursday work? how about Friday?” email costs a day. Connect the follow-up to each agent’s calendar so the buyer picks a slot, and send reminders the day before and an hour before. It cuts down on no-shows, and agents get their evenings back.
Where AI earns its keep
| Good use of AI | Keep it away from |
|---|---|
| Drafting a reply that answers the buyer’s actual question | Negotiating or discussing price |
| Summarizing a long thread before an agent calls | Promising availability, dates or terms |
| Pulling budget, area and bedrooms out of a free-text enquiry into CRM fields | Sending anything without an agent’s approval, at least for the first few weeks |
The setup we recommend is that AI drafts and the agent approves with one tap. Every message should be clear about who it’s from.
Keep in touch without nagging
Most buyers aren’t ready on day one. A simple sequence keeps you on their radar: new listings that match what they asked about, a short check-in after a week, a local market note after a month. The important part is stopping the sequence the moment they reply, so a person picks up the conversation instead of a bot sending another “just checking in”.
Measure the parts that leak
| Number | What it tells you |
|---|---|
| Time to first reply, by source and by agent | Where leads wait, and for whom |
| Enquiries that become viewings | Whether the first reply and booking work |
| Viewings that become offers | Lead quality and agent follow-through |
| Clients per source | Which portals and ads are worth paying for |
If you’re still choosing a CRM, we compared the two we use most in HubSpot vs GoHighLevel. And our real estate page shows the other work we do for property teams.
Mistakes we’d avoid
Sources
Questions we often hear
Won’t automated replies feel robotic?
The robotic ones are the generic “we’ll be in touch” messages. A reply that names the property, uses the agent’s name and hands over to a person quickly doesn’t feel automated at all.
GoHighLevel or HubSpot for a real estate team?
GoHighLevel suits teams that live on fast text follow-up and self-booked viewings. HubSpot suits larger brokerages that care more about reporting. We compare the two in our HubSpot vs GoHighLevel article.
Can you connect our listing portals and website forms?
Yes. We route enquiries from portals, web forms, ads and texts into one CRM, so nothing waits in a separate inbox.
