Homies Research.
Rigorous, practical work on the systems, standards, and models shaping real estate. Built for the people responsible for making AI useful—and accountable.
Governance
Policy that can be implemented
Benchmarks
Real jobs, shared harnesses
Open evidence
Claims linked to primary sources
Latest
Featured research
Written for directors, brokers, technical leaders, and builders.
The Complete Real Estate AI Integration Stack
An 84-system map of realtor CRM, MLS, IDX, email, voice, social, transaction and website integrations—with official API links and honest access status.
What Is an AI Harness? The Realtor’s Guide
A deep, practical guide to the operating layer that turns an AI model into a supervised real estate worker—with tools, memory, permissions, and audit trails.
How to Build a Real Estate AI Email Agent
The Gmail and Microsoft Graph architecture, CRM sync, prompt-injection controls, deliverability rules, costs, and build-versus-buy planner.
How to Build a Real Estate AI Text Messaging Agent
A technical blueprint for harness-to-CRM-to-Twilio messaging, consent, deliverability, webhooks, costs, and supervised two-way conversations.
How to Build a Real Estate AI Voice Agent
The complete 2026 stack, cost model, CRM architecture, compliance plan, and build-versus-buy calculator.
The Real Estate Social Media AI Agent Stack
A technical blueprint for AI clipping, digital twins, multi-platform posting, DM automation, CRM capture, and human approval.
The RAILS Framework
Interactive NAR/RPR adoption data and a capability-without-custody architecture for safe, compliant agentic AI in real estate.
HomieBench v3
Kimi K3, Claude Fable 5, GPT-5.6 Sol, and five more models ranked across 100 realtor workflows and completed-job unit economics.