Structured system prompts for agent and brokerage AI assistants, with fair-housing-aware guardrails.

A practical prompt pack for people building AI assistants for real estate agents and small brokerages. Every prompt uses the same structure: role, hard rules, tool-use guidance, refusal and escalation rules, and a JSON output schema your code can validate. Prompt-injection defenses are written into each one.

The 20 agents: buyer lead qualifier, seller lead brief, listing inquiry responder (AI-disclosed), showing scheduler, fair-housing-aware listing description writer, human-approved nurture drafter, transaction deadline tracker with wire-fraud flags, open house follow-up, market update summarizer, agent inbox triage, and review request drafter, CMA prep assistant, buyer consultation brief, offer comparison summary, rental inquiry pre-screen (AI-disclosed), post-close check-in drafter, disclosure checklist tracker, social listing post writer, vendor and inspector scheduling coordinator, and agent recruiting responder.

What's included:

PDF guide plus the editable Markdown source

Real estate guardrails checklist: fair housing and steering tests, no valuations or legal/lending advice, wire-fraud rules, lockbox-code handling, seller confidentiality, SMS consent, MLS data use

A step-by-step method for adapting each prompt

Who it's for: AI/automation freelancers and agencies building for real estate clients, and tech-savvy agents.

Requirements: any LLM that follows system prompts; JSON mode recommended.

What it's not: code, a plug-and-play agent, or legal advice. Fill in placeholders, connect your own tools, test on your own data, and review with the broker.