Build the brain
Write the brokerage rules, asset focus, scoring logic, and deal assumptions into plain markdown files.
Flo Agents / Commercial real estate
Owner intelligence for brokers who live on call lists.
Commercial real estate teams do not need AI magic first. They need clean owner context, better contact visibility, and a trusted way to decide who is worth calling next.
owner-flo / ranked-call-list.csv
Top owners to call
01
Harbor Ridge Holdings
Retail strip
Owned 11 years, clean contact record, no recent call logged.
89
02
Sunset Hospitality Group
Limited-service hotel
Financing note needs review, last contact older than 18 months.
84
03
Maple Court Partners
Apartment complex
Long hold period, verified email, missing working cell.
78
The wedge
Start with a brain, then build the tool on top.
The knowledge base holds the rules: asset focus, what makes an owner interesting, what contact data matters, how underwriting assumptions should be challenged, and what the team never wants AI to guess.
Brain files
- properties.md
- owner-signals.md
- call-list-rules.md
- underwriting-notes.md
- open-loops.md
First tool
Ranked owner call lists
Import a sample export, apply visible scoring rules, and produce the next owners to call with reasons the broker can inspect.
Proof-of-concept path
Import sample data
Use a redacted or fake CSV first: owner, property, asset class, years owned, financing note, contact status, last touch.
Rank the call list
Score each owner with visible rules, then show the broker why each person is worth calling now.
What comes later
Underwriting and CRM prompts get better once the context is stable.
The call-list builder is the first visible win. After that, the same brain can support underwriting assumption checks, deal-room summaries, CRM questions, and reusable broker playbooks.
Data posture
Start with fake or redacted data.
A public demo should never use a real brokerage export. Prove the shape with safe sample rows first, then decide what sources are approved before touching live owner contact data.