Real Estate Marketing Agents
Implementation fact Real business: implementation facts, no outcome metrics
A multi-agent system that runs my real estate marketing day: social posts, a weekly ad strategy, lead scoring and follow-up drafts, all tuned to the Baltimore market and kept inside housing-ad rules.
- Status
- Code on GitHub
- Relevant to
- AI and agents, GTM Systems
Results · Real business: implementation facts, no outcome metrics
- Implementation factA fact about what was built and tested, not a business outcome.lead score from contact quality, timeline, motivation, pre-qualification and sourceImplementation fact
- Implementation factA fact about what was built and tested, not a business outcome.in any ad, per the Housing Special Ad CategoryImplementation fact
The problem
A solo agent has to market every day, follow up with every lead and read ad results, on top of actually selling homes. Housing ads also come with fair-housing limits on how you can target.
What I built
- An orchestrator that runs specialist agents on a content calendar: content, ad strategy, lead scoring, follow-up, analytics and Baltimore market context.
- Bilingual output: posts and follow-ups are written in Spanish when the audience or the lead is Spanish-speaking.
- Ad strategy that stays inside Meta’s Housing Special Ad Category, and a compliance footer on every post.
- Draft mode by default: each day’s output lands in a review file before anything is posted.
Limitations
- Analytics use mock data until a real Meta access token is configured.
Where it came from
Scored leads come from my CRM, and every post and follow-up starts as a draft I review. Read the AI CRM case study.
Stack
Python, Anthropic API, Meta Graph API