You know the feeling. You're in a listing appointment and three enquiries land in your inbox. By the time you're back in the car, two of those buyers have already heard from someone else. That gap, between knowing you should reply and actually being able to, is where commission goes to die. This week's research is almost entirely about closing that gap.
This week's play
AI first-draft for every new enquiry
The average agent takes around 15 hours to reply to a new lead. Buyers go with whoever calls first, 78% of the time. This play gets a personalised reply and two follow-ups drafted in under two minutes, so you review and send rather than write from scratch.
- Save the prompt below in your AI tool of choice (ChatGPT, Claude, or Gemini all work). Fill in your business details once.
- When a new enquiry arrives, paste the full message into the chat and run the prompt. You get a summary, a reply, and two follow-ups in one output.
- Review for 30 seconds, check the name and any price references, then send Reply 1 from your normal email. Schedule Follow-up 1 for two days out and Follow-up 2 for seven days out using your email client's schedule-send feature.
- Track it for four weeks: time to first reply, meetings booked, and deals progressed. One narrow workflow, measured properly, tells you whether to go further.
You are my sales assistant for a residential real estate agency. Business: [describe in 2-3 sentences: your market, property types, typical price range]. Tone: professional but warm, plain language, no jargon. Typical next step I offer: [e.g., a call to discuss their search, a viewing, a free appraisal]. For the enquiry I paste below, do the following: 1. Extract and list: name, contact details, what they are looking for, budget or timeline if mentioned. 2. Draft: - Reply 1 (send now): under 180 words, acknowledges their specific request, asks up to 3 qualifying questions, offers my next step. - Follow-up 1 (send in 2 days): short, references the original enquiry, one or two questions, repeats the next-step offer. - Follow-up 2 (send in 7 days): brief final check-in, leaves the door open. 3. Output as: Summary / Reply 1 / Follow-up 1 / Follow-up 2. Here is the enquiry: [paste the full enquiry text here]
Speed-to-Lead and the Database
The follow-up problem is real and measurable, and AI is now doing the work agents can't get to fast enough.
- That buyer who enquired at 11pm last night, the one you'll call first thing tomorrow, has probably already booked a viewing with whoever picked up. monday.com built an AI intake assistant called Amanda that answers inbound sales requests, asks qualifying questions, and books the meeting, dropping response time from 24 hours to under 2 minutes. The same company reported 2.5x conversion for leads handled this way versus a control group.For your business Set up an AI-drafted first response for every portal enquiry, even if you review and send it manually. Buyers who get a reply in minutes are far more likely to stay yours. That is the listing pipeline you protect.
- Your dead leads are not dead, they are just waiting for someone to follow up. Research prep that used to take a rep one to two weeks, pulling account history, writing a personalised outreach sequence, is now being done in about five minutes by AI tools doing the same job. That means your cold database from six months ago is workable again without a full day of admin.For your business Pick 20 contacts from your CRM who went quiet after an appraisal or open home. Paste their details into ChatGPT or Claude and ask it to draft a short, personalised check-in for each. Send five today. The database is your goldmine and it is sitting there.
- Anthropic's Claude is now embedded inside HubSpot, Google Workspace, Microsoft 365, and DocuSign, reported by Forbes on June 19. In plain terms, the AI that drafts your follow-up can now also pull from your contact records, check your calendar, and prep a contract summary, all inside tools you already use.For your business If your CRM is HubSpot, check what AI features are live in your account this week. Even one automated follow-up sequence saves you the 15-hour average response lag that costs agents buyers.
Marketing Your Listings
AI is now handling the full content loop, from listing copy to social assets to branded email campaigns, faster than any one-person office can manage manually.
- You have a listing going live Thursday and you need copy, a social post, and an email to your database. ChatGPT can now draft the full email, pull in images from a connected library, preview the layout on desktop and mobile, and hand you a link to review, all in one guided flow. What used to be three separate tasks across two tools is now one conversation.For your business Write your listing brief once, then ask your AI tool to produce the email, the social caption, and the portal description from the same brief. You get consistent messaging across every channel and your Thursday night back.
- Buyers are increasingly finding properties and agents through AI assistants, not just portals. Google Analytics 4 now tracks traffic arriving from ChatGPT, Gemini, and Claude in a separate channel called AI Assistant. Agents who have a clear, factual presence online are starting to show up in those answers. Agents who do not are invisible.For your business Check whether your website or profile pages are being found via AI-driven search. If you do not know, that is the answer. Being the agent an AI assistant names when a buyer asks is the new version of ranking on page one.
Transaction Admin and the Back Office
Contract-to-close admin is the part of the job nobody talks about at awards nights, and AI is quietly eating through it.
- Your transaction coordinator is manually re-keying every price, date, contingency, and party name from the executed contract into your back-office system. ReBillion's AI back-office layer, documented as live in June 2026, reads the contract, extracts every field, flags missing signatures, auto-names and files documents, and maps inbound email to the correct deal, all with an audit trail. The TC becomes a reviewer, not a typist.For your business If your brokerage runs on SkySlope, Dotloop, or Brokermint, look at what AI contract-review tools now connect to those systems. One fewer manual step per file across a month of closings is a meaningful time saving for a small team.
- James Dwiggins, CEO of NextHome, posted this month about using Claude to read Dotloop activity logs, identify which files changed since his last review, and produce a prioritised list for his attention. He also flagged using AI to pre-check listing descriptions and agent communications for compliance issues before they go live.For your business Before your next listing goes live, paste the description into Claude or ChatGPT and ask it to flag anything that could create a compliance issue in your market. Catching a problem before publication is far cheaper than after.
- Missed calls after hours are missed listings. AI voice tools now answer prospect calls at any hour, ask BANT-style qualifying questions, and book the showing directly into the agent's calendar. The lead that rings at 9pm on a Sunday no longer goes to voicemail and then to your competitor.For your business If you are not capturing after-hours enquiries, work out how many calls you miss in a typical week. Even one missed listing appointment per month has a real GCI cost. AI phone intake is now a practical option, not a future one.
From the feed
One thing to do this week: save the prompt above and run it on the next enquiry that comes in. See how long the review takes and whether the reply is good enough to send. If you also want to be the agent AI assistants name when a buyer asks who to call in your suburb, that is exactly what Rollo builds. joinrollo.com.
Sources: Growth Unhinged - AI-native GTM at public company scale · Forbes - AI slipped into your business tools · Marketing with AI Weekly Roundup · Ignite Visibility - Digital Marketing News · ReBillion - AI Real Estate Back Office · James Dwiggins - LinkedIn · CallSphere - AI for Real Estate · Prasanna Kumar - LinkedIn