What's Inside
How Realtors Work — Your full production workflow broken down by business area, with AI opportunities mapped to each stage. 🗺️
This Week in AI — Frontier labs argued about slowing down while the tools you actually use kept getting quietly folded into your everyday software. 🗞️
Top Picks — An AI receptionist, a CRM copilot, an AI video B-roll generator, a lead-scoring agent, a proposal automation, and a writing-style trick. 🔧
Deep Dives:
📞 ElevenLabs Reception — An AI receptionist that answers after-hours calls and books showings for you
🤝 Salesforce in Claude — A CRM copilot for account research, call prep, and pipeline updates
🎬 Riverside AI B-roll — Generate short video clips for listing and social content right inside your editor
🎯 Nex — An AI agent that scores your leads against your own closed deals and drafts win-back emails
📝 Zapier Call-to-Proposal Automation — Turn meeting notes into a drafted follow-up or proposal automatically
✍️ Compound Writing — Turn your own edits into reusable AI writing rules so drafts finally sound like you
Mini Skill — Turn a handful of property bullet points into a polished MLS listing description in under 10 minutes. ✍️
How Realtors Work: An AI Opportunity Map
Lead Generation
Your pipeline starts with sphere outreach, expired listings, FSBOs, online leads, door knocking, and social media, usually run in parallel with no single system tying them together. Most agents lose time re-researching the same prospect across multiple sources before ever making contact. The friction point is almost always the front end: figuring out who to contact and what to say first, before any actual selling happens.
Lead Nurturing
Once a lead exists, it needs CRM follow-up, email sequences, text and call cadences, and periodic market updates to stay warm until it's ready to convert. Most agents nurture inconsistently because the manual work of writing and timing every touchpoint competes with everything else on their plate. Leads that don't convert in the first 30 days quietly go cold here, not because they weren't good leads, but because nobody kept in touch.
Buyer Representation
This covers the buyer consultation, property search, showing coordination, offer writing, negotiation, and transaction management through closing. The time sink is usually coordination: chasing showing confirmations, comparing properties across client preferences, and re-explaining contract terms client by client. Offer writing and negotiation prep also eat hours that could be spent with clients instead of paperwork.
Seller Representation
Seller work spans the listing presentation, pricing strategy, listing prep, photography coordination, marketing, open houses, offer review, negotiation, and transaction management. Pricing strategy and the listing presentation itself take the most prep time, since they require pulling and packaging real market data persuasively. Coordinating vendors (photographers, stagers, contractors) around a seller's timeline is a second common bottleneck.
Marketing and Content
This includes social media, video, email newsletters, community content, and personal branding, all of which compete for time against actual transactions. Most agents either under-post because content creation feels like a full-time job on its own, or they outsource it and lose their authentic voice in the process. Consistency, not creativity, is usually the real gap.
Administrative and Operations
Contract management, deadline tracking, vendor coordination, commission tracking, and compliance sit behind every deal and rarely get automated. A missed deadline or an untracked commission split can cost real money, so this area carries outsized risk for how little attention it usually gets. Most of this work is repetitive enough to template, but few agents take the time to build the templates.
Client Experience and Retention
Closing gifts, post-close follow-up, referral cultivation, and review or testimonial requests determine whether a client becomes a repeat client or a one-time transaction. This is the stage most agents skip entirely once the commission check clears, even though it's the cheapest way to generate future business. A simple, consistent follow-up cadence after closing is one of the highest-leverage, lowest-effort habits in the business.
Professional Development and Recruiting
For team leads and brokers, this covers agent training, onboarding, and recruiting, on top of running your own production. Keeping training materials current and personalizing onboarding for each new agent takes real time that's easy to deprioritize in favor of closing deals. This is often the first area to go stale once an agent starts building a team.
This Week in AI — What It Means for Realtors
This week's AI headlines were dominated by a very public argument among frontier AI labs about whether to slow down development, with Anthropic's CEO calling for a coordinated pause and OpenAI, Meta, and Microsoft all weighing in with different answers. Washington and Beijing both waved it off within a day. None of that changes your Monday, but it's worth noticing what was happening underneath the noise: the same week everyone argued about pacing AI, the tools you actually touch kept getting more capable and more woven into the software you already use.
That's the real pattern this week. Anthropic folded its separate Cowork feature back into the main Claude app and added Docs and Slides. Salesforce built an AI layer directly into its CRM instead of shipping another standalone app. Apple finally shipped a rebuilt Siri with real on-screen and cross-app awareness. None of these are flashy new products you'd have to go find and sign up for, they're AI quietly becoming a built-in feature of software you may already pay for.
The takeaway for a working Realtor: the highest-leverage move right now isn't chasing every new AI app that launches, it's checking whether the tools already in your stack (your CRM, your phone system, your video editor) have quietly gained an AI layer worth turning on. This week's picks below are a mix of both: a few standalone tools worth trying, and a couple of features that might already be sitting inside something you use.
Top Picks This Week
ElevenLabs Reception
What it is: An AI receptionist that answers phone calls, handles FAQs, and books appointments in a natural-sounding voice.
Realtor use case: Catch after-hours calls from buyers and sellers instead of losing them to voicemail, and let it book showings directly onto your calendar.
Why it matters: Missed calls are missed leads, and this handles the ones that come in after you've clocked out.
Salesforce in Claude
What it is: An AI layer built into Salesforce with dozens of skills for account research, call prep, pipeline reviews, and CRM updates.
Realtor use case: If your brokerage runs Salesforce, pull a client's full history before a call or update records in plain language instead of manual data entry.
Why it matters: It's a preview of where every major CRM is heading, worth knowing about even if you're not on Salesforce yet.
Riverside AI B-roll
What it is: AI-generated video clips built directly into Riverside's editor, turning a text prompt into an ~8-second clip.
Realtor use case: Fill gaps in listing walkthroughs or social reels with transition and lifestyle shots without a second camera setup.
Why it matters: Video sells listings, and this removes one more excuse for skipping it.
Nex
What it is: An AI agent that scores leads against your own closed-won deals instead of a generic checklist, then drafts win-back emails.
Realtor use case: Stop guessing which cold leads in your database are worth another call, and get a first-draft re-engagement email for the ones that are.
Why it matters: Most agents have a graveyard of old leads they never revisit. This gives you a reason to.
Zapier Call-to-Proposal Automation
What it is: A workflow that takes your meeting notes, has AI pull out key details, and drafts a follow-up email or proposal from a template.
Realtor use case: After a listing appointment or buyer consultation, get a drafted recap email or next-steps summary waiting for your review.
Why it matters: It closes the gap between "great meeting" and "actually followed up," which is where a lot of business quietly slips away.
Compound Writing
What it is: A prompt technique that turns your edits to an AI draft into reusable style rules for next time.
Realtor use case: Stop re-explaining your voice every time you ask ChatGPT or Claude to write a listing description or client email.
Why it matters: It's free, takes ten minutes to set up, and makes every future AI draft sound a little more like you.
Deep Dives
ElevenLabs Reception
Reception is an AI phone receptionist built by ElevenLabs, the company known for realistic AI voice technology. It answers incoming calls, handles common questions, and can book appointments straight onto a connected calendar, all in a voice that sounds like a real person rather than an automated menu.
For a Realtor, the clearest use is after-hours and overflow coverage. A buyer calling about a listing at 8pm on a Saturday, or a seller trying to reach you during a showing, gets a real conversation and a booked callback or showing instead of a voicemail they may never leave. It's also useful for agents who don't want to hire a part-time assistant just to cover phones.
Setup runs through ElevenLabs' platform: you connect a phone number, write or approve a script for common questions, and configure how it hands off to your calendar or CRM. This is a paid product tied to ElevenLabs' subscription tiers, and initial setup for a basic call-and-book flow takes roughly 30 to 60 minutes, longer if you want it to sound highly specific to your business.
The caveat: this is a newer product category, so test it with your own scripts before trusting it with real leads, and expect to spend some time tuning the calendar and CRM handoff so bookings land where you actually want them.
Salesforce in Claude
Salesforce in Claude embeds Anthropic's Claude directly into the Salesforce CRM, giving it built-in skills for tasks like account research, call prep, pipeline reviews, and updating records using plain-language instructions instead of manual data entry.
The real estate parallel is direct even if you don't personally use Salesforce: this is what an AI-native CRM looks like. Instead of opening five tabs to pull a client's history before a call, you'd ask for a summary. Instead of manually logging notes after a showing, you'd tell the system what happened and have it update the right fields. If your brokerage or team uses Salesforce, this is available now.
It's currently in beta, and availability depends on your Salesforce and Claude licensing through your brokerage's IT or admin team, so this isn't a five-minute personal setup. For independent agents on other CRMs (Follow Up Boss, kvCORE, Wise Agent, and similar), this is worth watching rather than adopting today, since it signals where those platforms are likely headed next.
The limitation is obvious: it only helps if you're on Salesforce. If you're not, ask your broker whether something similar is on the roadmap for your actual CRM.
Riverside AI B-roll
Riverside, a video and podcast recording platform, added AI-generated B-roll directly inside its editor. Type a text description of the shot you want, and it generates a short clip (around 8 seconds, built on Google's Veo 3 model) that you can drop straight onto your timeline alongside your real footage.
For listing videos and social content, this fills the gaps that usually require a second camera trip or paid stock footage: a neighborhood establishing shot, a lifestyle cutaway, a transition between rooms. It won't replace your primary walkthrough footage, but it can round out a video that would otherwise feel thin.
It's available to Riverside's Pro+ subscribers and uses AI credits per generation, so there's a marginal cost per clip rather than a flat fee. If you're already recording listing videos or client testimonials in Riverside, this adds on top of your existing subscription with no separate signup.
The catch: generated clips can still look slightly synthetic up close, so use them for supporting shots, not the hero footage of the actual property, and budget your credits since generation isn't unlimited.
Nex
Nex is an AI agent built specifically for lead scoring and re-engagement. Instead of ranking leads against a generic checklist, it scores them against the pattern of your own closed-won deals, then drafts win-back emails for the cold accounts most likely to convert based on that pattern.
The use case for a Realtor is the pile of old leads sitting untouched in most CRMs: past showings that went nowhere, sphere contacts who went quiet, online leads that never returned a call. Nex ranks that list based on what actually worked for you before, then gives you a drafted email to personalize and send, instead of you starting from a blank page or ignoring the list entirely.
Setup requires connecting your closed-deal data so the scoring has something to learn from, which is a real onboarding step, not a five-minute signup. Pricing isn't published, so request a demo or quote before committing rather than assuming it fits a solo agent's budget.
The limitation: it's only as good as the closed-deal data you feed it, and without published pricing, you'll want to confirm cost and contract terms upfront.
Zapier Call-to-Proposal Automation
This is a workflow, not a standalone product: a Zapier automation that takes your meeting or call notes, uses AI to extract the key details, fills them into a document template, and drafts a follow-up email or proposal for your review.
Adapted for real estate, this fits neatly after a listing appointment or buyer consultation. Instead of remembering to write up a recap email or draft next steps later that night, the workflow does a first pass automatically, pulling from your notes into a template you've already built (a recap email, a next-steps summary, or a simple proposal document).
Setup requires a Zapier account (the free tier works for testing), a notes app you already use, and a document template with placeholder fields. The Rundown's guide walks through building this in plain language using Zapier's Copilot, and most agents could have a basic version running in 15 to 20 minutes.
Treat every output as a draft, never send-ready. In testing, the AI has left vague placeholder language or approximate numbers where specifics belonged, so always read before sending. It works best on repeatable, templated documents rather than one-off, highly custom communications.
Compound Writing
Compound Writing is a prompt technique (with a free, open-source companion plugin) built around a simple idea: every time you edit an AI-written draft, you can extract what changed into a reusable rule, so the next draft comes back closer to your actual voice.
If you're using ChatGPT or Claude to draft listing descriptions, client emails, or social captions, this solves the recurring annoyance of re-explaining your tone and preferences every single session. After you edit a draft, you ask the AI to compare its version with your edited version and pull out only the reusable lessons, not one-off factual corrections, then save those as a short instruction file you paste in at the start of future drafts.
It's completely free and works with any major LLM, no signup or subscription involved. Setup is just the habit of running the comparison prompt after you edit something and keeping the resulting rules file somewhere handy.
The limitation is that it's manual, not automatic memory. You have to remember to run it and reuse the saved rules yourself; if you skip that step for a few weeks, you lose the compounding benefit that gives it its name.
This Week's AI Mini Skill — 10 Minutes, Real Results
This week's skill: turning a handful of property bullet points into a polished MLS listing description, using ChatGPT, Claude, or Gemini (free tier works fine for this).
Open ChatGPT, Claude, or Gemini in your browser.
Jot down the property's key details as quick bullet points: bed/bath count, square footage or lot size, standout features, neighborhood highlights, and anything else you'd normally lead with.
Paste the prompt below into the chat, filling in your bullet points where indicated.
Read the draft against your actual listing sheet, since AI can occasionally invent a detail that sounds plausible but isn't true. Correct anything inaccurate.
Rewrite the opening line in your own words so it doesn't read like every other AI-generated listing, then paste the rest into your MLS entry, website, or flyer.
Copy-paste prompt:
You are a real estate copywriter. Turn these property details into a compelling MLS listing description under 200 words. Highlight the top 3 selling points based on what's most unique or valuable about this property. Avoid cliches like "must see" or "won't last." Write in an inviting but factual tone, with no exaggeration or invented details.
Property details:
- Bedrooms/bathrooms: [fill in]
- Square footage/lot size: [fill in]
- Key features: [fill in, e.g. renovated kitchen, view, finished basement]
- Neighborhood highlights: [fill in]
- Target buyer or price point (if relevant): [fill in]What to do with the output: treat it as a strong first draft, not a final copy-paste. Verify every fact against your listing sheet, swap the opening sentence for something in your own voice, and then it's ready for the MLS, your website, or a listing flyer.
The AI Edge for Realtors is published weekly. If you found this useful, forward it to another agent in your office.