82% of real estate agents now use AI. Most are paying $300-500/month for it.
A February 2026 NAR survey found that 82% of US real estate agents now use AI tools in their daily work, up from 68% in 2025 and roughly 15% in 2023. The category exploded over 18 months. With that came the SaaS bills.
Lofty costs $300-500 per agent per month. Structurely runs $350+. Follow Up Boss $300+. SmartZip enterprise tier $1,000+. The cheap end (Listing Copy AI, PhotoAIVideo, Write.Homes) starts at $20-99/month — recurring forever.
If you close 6-30 transactions a year as an individual agent, the math on $400/month per seat is real. That's $4,800/year — equivalent to commission on a $320K property at 1.5% commission. You don't need a calculator to figure out whether the AI tool is paying for itself.
This post breaks down what those real estate AI tools are actually doing, why the value lives in the prompt patterns (not the UI), and what a $49 lifetime pack of those patterns looks like.
What Lofty, Structurely, and Follow Up Boss actually do
Strip the marketing and the agent-AI category converges on the same feature set:
- MLS-compliant listing descriptions. Take property details and generate a description that satisfies MLS character limits, Fair Housing requirements, and broker standards.
- Buyer / seller email drafting. Convert raw notes into client-ready emails for showings, offers, counter-offers, and updates.
- Comp analysis from raw MLS data. Take comp pulls and produce a structured pricing recommendation.
- Lead triage and nurture. Score and sequence incoming leads from Zillow, Realtor.com, and Facebook.
- Marketing content. Generate Instagram captions, video reel scripts, neighborhood market reports.
- Inspection report summaries. Convert technical inspection reports into buyer-friendly summaries.
Everything else — the CRM, the IDX integration, the website hosting — is SaaS plumbing. The actual AI value lives in how the tool prompts the underlying LLM to handle each of these tasks.
What makes a real estate AI prompt actually work?
The Real Estate Agent + Claude Code Prompt Pack uses seven structural patterns. Two are especially load-bearing for real estate work because of the regulatory environment.
The Fair Housing constraint is the highest-leverage prompt design decision. Every listing-description prompt must produce output that describes the PROPERTY, not the IDEAL OCCUPANT. Phrases like "perfect for a young family" or "great walkable neighborhood for active people" are Fair Housing violations under the Federal Fair Housing Act and most state equivalents. AI tools that don't bake in these constraints up front will eventually generate output that exposes the agent to a complaint.
The pack's listing-description prompt includes an explicit Fair Housing compliance check that flags any phrase potentially violating the rules. This is the difference between a tool that helps you and a tool that gets you sued.
The MLS compliance constraint is the second. Most MLS systems have character limits (typically 1,500-2,500 chars), prohibited terms (specific marketing claims), and required fields. The pack's prompts include explicit MLS-aware formatting that ships output ready to paste.
The other five patterns (role anchoring, context fences, output contract, failure modes, iteration anchors) apply the same way they apply across the seven-patterns framework.
Three example prompts from the Real Estate Pack
Example 1: MLS-compliant listing description from raw property details
The prompt takes property attributes (address, beds/baths, sqft, lot, year built, recent updates, standout features) and produces a listing description that:
- Opens with a HOOK first line that captures the property's defining quality WITHOUT prohibited Fair Housing terms (no "perfect for", "ideal for", "great family neighborhood").
- Presents factual property attributes in skimmable order.
- Describes standout features in concrete terms ("white oak floors throughout main level" not "luxurious flooring").
- Describes the neighborhood BY THE PROPERTY (proximity, view, amenities in neutral terms), not BY THE DEMOGRAPHICS or who lives there.
- Closes with a CTA that invites action without pressure ("Schedule a private tour" not "Won't last!" which is also a regulatory tripwire in some states).
- Stays within the agent's MLS character limit.
The prompt ALSO outputs a Fair Housing compliance check that flags any phrase that could be problematic. This is the part the cheap SaaS tools skip and the part that protects the agent's license.
Example 2: Comp analysis from raw MLS data
The prompt takes comp pulls (3-7 recent comps with sold price, list price, DOM, bd/ba, sqft, lot, year, condition notes) and produces:
- A clean comp grid with the standard adjustments (size, condition, location, time).
- The adjustment methodology — how each comp adjusts to the subject (typically +/- $/sqft for size, $/major-system for condition, market-trend factor for time).
- A recommended price range with reasoning.
- The price the seller will likely WANT to hear vs the price the market will actually pay (this is the gap that loses listings).
- Strategic recommendation — price for the market or price for negotiation room.
- Seller-conversation talking points for the listing appointment.
Most agents do this analysis in their head. Writing it down — even with AI assistance — is what turns the listing appointment into a structured conversation instead of a vibe check.
Example 3: Buyer offer-strategy memo for a competitive bid
The prompt produces an offer-strategy memo with the buyer's maximum budget, the comp range, the market temperature, and (crucially) the buyer's flexibility on closing date and contingencies as a relative advantage.
The memo includes an HONEST risk section. A buyer who waives an inspection contingency on a 60-year-old house is taking real risk. The memo says so. This is the conversation most agents shorthand or skip — and the conversation that turns into a regret-driven listing six months after closing.
The pack includes a 1-page risk acknowledgment the buyer can sign confirming the trade-offs they understand. This is the file documentation that protects you on the back end.
What the pack contains
The Real Estate Agent + Claude Code Prompt Pack ships v1.0 with 8 starter prompts (2 per role) at full quality, expanding to all 50 prompts in v1.1 within 4 weeks. Your $49 unlocks both.
Listing Agent: MLS-compliant listing descriptions, comp analysis, inspection-report summaries for buyers, seller weekly update emails, "why your home didn't sell" recovery scripts, listing-price-adjustment letters, multiple-offer presentation frameworks, contingent-offer evaluation, contract-review checklists for sellers, addendum drafting, repair-request responses, appraisal-shortfall conversations, post-closing wrap-up.
Buyer Agent: buyer intake / preferences questionnaire that surfaces real motivation, offer-strategy memos for competitive bids, comp-property write-ups from MLS data, tour-day talking points, mortgage-broker referral handoff, post-tour debriefs, inspection-day talking points, due-diligence checklists, post-offer-acceptance roadmap, appraisal-day prep, final-walkthrough checklist, post-closing closeout.
Lead Gen / Marketing: monthly neighborhood market report emails, open-house follow-up email sequences, IG carousel scripts, video reel scripts (30-day market update, just-sold, neighborhood spotlight, first-time-buyer education), past-client check-in cadence, listing-just-listed announcements (3 channels), open-house promotion, FSBO outreach, expired-listing outreach, geographic-farm postcard campaign, social-proof testimonial requests, referral-thank-you campaign.
Brokerage / Team Lead: new-agent 30-day onboarding curriculum, recruiting-an-agent pitch outlines, transaction-file audit checklists, weekly team standup notes, broker-of-record letters, agent performance review framework, team marketing calendar, listing-coordinator workflow, transaction-coordinator workflow, vendor evaluation, commission-dispute resolution, succession-planning conversations.
Full pack at clskillshub.com/pack/real-estate.
Why $49 lifetime vs $300-500/month
Lofty, Structurely, Follow Up Boss, and SmartZip charge monthly because they bundle: AI capabilities + CRM + IDX website + texting + SMS automation + analytics. That's a lot. If you need a full real-estate-business operating system, those tools earn their price.
The pack ships just the prompt patterns — the part that makes the AI capabilities good. You bring your own CRM, your own MLS access, your own Claude or ChatGPT account. The trade is convenience for ownership.
For agents who already have a CRM they love (or who use IDX hosting separately), paying $300/month for ANOTHER CRM just to get the AI features is poor math. The pack lets you keep your existing stack and add the AI structure on top.
Who should buy this
Individual agents (6-30 transactions/year), team leads of 3-10 agents, and brokerage marketing leads who:
- Are already using ChatGPT, Claude, or Gemini for listing copy / emails / market reports ad hoc
- Want structured prompts that produce Fair-Housing-safe output consistently
- Already have a CRM and don't want to switch just to get AI features
- Are comfortable using Claude Code (CLI) or pasting prompts into Claude.ai
- Run enough volume to justify owning the patterns vs renting
NOT for agents who need: full real-estate operating system with CRM + IDX + AI bundled, hand-holding through every campaign setup, or a vendor's customer success team. Those needs are real; the pack solves a different problem.
How to use the pack
1. Copy-paste into Claude.ai or another LLM. Pick the prompt, paste, fill in bracketed inputs.
2. Install as Claude Code skills. Drop skills-install/real-estate/ into ~/.claude/skills/, restart Claude Code, type @real-estate in any session.
3. Searchable reference. Ctrl+F for "MLS", "comp", "open house", "FSBO".
Related reading
- The 7 Patterns Behind High-Performance Claude Prompts — the prompt framework every prompt uses
- Harvey AI Alternative for Lawyers — same approach for legal
- Pricing page — all 7 industry packs
- Real Estate Pack on clskillshub.com
FAQ
How does this compare to Lofty? Lofty is a managed real-estate operating system at $300-500/month per agent. The Real Estate Pack is a one-time $49 purchase of the prompt patterns that make AI tools useful for real-estate work. Lofty's CRM, IDX, and analytics are not in the pack — but the prompt structure that does the AI work is.
Will the listing descriptions pass Fair Housing review? Each listing-description prompt produces output that is Fair-Housing-compliant by design (describes the property, not the occupant), and outputs a compliance check that flags any phrase that could be problematic. You remain responsible for the final review — every output should be read by the agent before posting.
Does this work with my MLS? The prompts produce output in plain text. Your MLS character limits, prohibited terms, and required fields are inputs to the prompt — the prompt adjusts to fit your specific MLS's rules.
Why ship v1.0 with 8 prompts and not 50? To validate buyer signal on what scenarios matter most before committing to the remaining 42. Your $49 includes v1.1 (full 50 prompts) within 4 weeks, automatically, via the same unlock link.
Lifetime updates? Yes.
Refund policy? Digital product, all sales final. If something genuinely does not land for you, reply to your purchase email and I will add you to the full Skills Library (lifetime access) as a goodwill gesture.