How to read this landscape
Every AI tool a real estate agent could buy falls into one of seven functional categories. A team rarely needs more than three. The most common mistake is buying a "platform" that bundles all seven and then using only one of them at $400/month.
The seven categories: (1) Conversational/ISA agents, (2) CRM-embedded AI, (3) Content production AI, (4) Listing copy AI, (5) CMA/pricing AI, (6) Transaction document AI, (7) Custom agents / orchestration.
Before you pick anything, walk the 10 swimlane workflows in the Gugubrand library and identify which specific step is the bottleneck for your team. The right AI tool is the one that owns that step end-to-end — not the one with the prettiest landing page.
1. Conversational / ISA agents
What they do: Receive an inbound lead, run an LPMAMA-style conversation over SMS or voice, qualify, book an appointment, escalate to a human if needed. The core workflow this category replaces is Lead Capture (Workflow 01).
Who it is for: Any team paying for inbound leads (Zillow, Realtor.com, Facebook Lead Ads) where speed-to-lead is below 95% in 5 minutes.
Price band: $200–$800/month per agent seat for off-the-shelf SaaS. Custom AI agents (Gugubrand, in-house builds) typically $600–$3,000/month flat.
Where it breaks: Long-tail edge cases. AI handles the majority of routine inbound conversations cleanly; a meaningful share still need real human judgment. If your "AI ISA" has no escalation path, it will hallucinate confidently on the 15–30% that should be transferred.
2. CRM-embedded AI
What they do: Lead scoring, smart routing, drip-cadence optimization, conversation summarization inside your existing CRM (Follow Up Boss, BoomTown, Sierra Interactive, kvCORE).
Who it is for: Teams already on a major CRM that bills annually for the AI add-on. If your CRM does not have it natively, do not buy a second tool — fix the CRM choice.
Price band: Typically a $100–$300/month add-on to existing CRM seat licenses.
Where it breaks: Lead scoring is only as good as your historical close data. If you have under 100 closed deals in the CRM, the model is guessing. Use rule-based routing first; ML scoring later.
3. Content production AI (social, email, video)
What they do: Generate captions, post copy, video scripts, monthly newsletter content, hashtag suggestions. Some include scheduling and analytics; most do not.
Who it is for: Teams with a Marketing Coordinator producing 10+ hours of content per week.
Price band: $30–$120/month per user. Many include unlimited generation.
Where it breaks: Tone drift. AI-generated content sounds like AI-generated content unless your prompts include 5–10 examples of your actual past posts. Without that calibration, your social presence reads as generic.
4. Listing copy AI
What they do: Generate the MLS description, social caption, ad copy, and email blast copy from photo + MLS feed input.
Who it is for: Listing-heavy teams (2+ listings per month per agent). Lower-volume teams should hand-write descriptions and use the formula in our [Listing Description Formula](/real-estate/listing-description-formula) piece.
Price band: $20–$80/month per agent seat, or per-listing pricing ($2–$10).
Where it breaks: Fair Housing violations. Generic AI-generated copy will produce banned language unless explicitly constrained. Every listing-copy AI deployment needs a Fair Housing classifier or regex check before MLS upload — and human review for any flagged copy.
Fair Housing on AI-generated listing copy
Off-the-shelf listing-copy AI tools vary wildly in their Fair Housing guardrails. Before buying, ask the vendor: (1) what banned phrases are filtered, (2) is there a post-generation classifier, (3) does the output go to MLS without human review. If the answer to #3 is "yes," do not buy.
5. CMA and pricing AI
What they do: Pull comparable sales, apply adjustment math, generate a draft pricing band. Best of them produce a seller-ready presentation; weaker ones produce a CSV.
Who it is for: Listing agents and pricing-heavy teams. See our [CMA Playbook](/real-estate/cma-playbook) for the underlying framework.
Price band: $40–$150/month per agent seat.
Where it breaks: Condition assessment. AI cannot smell the basement. Do not trust photo-only condition tier calls — always walk the comp or pull MLS photos and human-verify.
6. Transaction document AI
What they do: Document checklist tracking, deadline alerts, e-signature workflow, contract analysis, broker-compliance file compilation.
Who it is for: Teams running their own transaction coordination (no external TC). If you outsource TC, this lives at the TC vendor — do not buy your own.
Price band: $50–$200/month per agent.
Where it breaks: Edge cases in state-specific contract language. A national tool may not know your state's repair contingency language; verify against your local board's standard forms.
7. Custom agents and orchestration
What it is: Custom AI agents built specifically for your team's workflows — typically by a partner like Gugubrand, or in-house if you have a technical co-founder. Often includes orchestration between multiple tools (e.g., new lead in CRM → AI ISA runs LPMAMA → if qualified, AI books appointment → if specific objection, escalates to Slack).
Who it is for: Teams of 4+ agents with >$2M annual GCI who have already adopted Categories 1–6 and are hitting orchestration limits.
Price band: $1,500–$10,000/month flat, depending on agent complexity and integration depth. Often replaces 1–2 admin or ISA hires.
Where it breaks: Vendor selection. If you cannot see the system prompts, escalation rules, and compliance guardrails, you are buying a black box. Pick partners who show their work.
The decision framework
Step 1 — Diagnose your bottleneck first, not the tool. Walk the [10 swimlane workflows](/real-estate). Identify the single step costing you the most hours or commissions per month.
Step 2 — Match the bottleneck to one of the seven categories. Most teams' bottleneck is in Category 1 (speed-to-lead), Category 3 (content), or Category 6 (TC). Pick one.
Step 3 — Trial the top two vendors in that category for 30 days. Measure: hours saved, leads contacted, appointments set. Anything else is marketing.
Step 4 — Add the next category only after the first is fully adopted. Most teams fail with AI by buying three tools at once and using none of them.
Step 5 — Custom agents come last, not first. Off-the-shelf wins the first $5K of monthly leverage. Custom wins the $50K and above.
What we deliberately left out
AI headshot generators (one-time tool, not a workflow). AI lead-magnet-page generators (not real-estate-specific). "Faceless" Instagram content (not legal for licensed agent marketing in most states). Generic ChatGPT subscriptions (useful as a side tool, not a system).
If a vendor pitches an "AI platform" that bundles all 7 categories, ask which one they're actually best at. Buy that one only. Add the rest from other vendors.