Roadmap
What this product looks like with CoStar firm-level data access.
CoStar is the firm-level data layer for commercial real estate. The current product runs on a stack of public-API alternatives (ATTOM / RentCast / HouseCanary / FRED) and a paste-driven comp-grid workflow. The architecture is built so CoStar slots in without rewriting the underwriting engine — the data_providers/costar.py stub class already implements the provider interface.
Data points: current vs CoStar
| Data point | Current (ATTOM / HC / RC) | With CoStar |
|---|---|---|
| Multifamily sales comps | Manually pasted (ATTOM gates comps behind paid tier) | Real-time, all transactions ≥ 5 units |
| Lease comps | Limited to active listings (RentCast) | All in-place leases including effective rent, concessions, lease terms |
| Property fundamentals | Public records (ATTOM) | Property-level rent rolls, T-12 financials, capex history |
| Submarket fundamentals | Manually researched | Full submarket forecasts: rent growth, occupancy, supply pipeline |
| Investment activity | Not available | Live trades, asking spreads, capital flows by market |
| Tenant data | Not available | Tenant credit ratings, lease maturity stacks |
| Building amenities | Listing photos only | Standardized amenity scoring, comparable analysis |
Specific memo improvements with CoStar
- Sales comp adjustment grid: instead of user-pasted comps, pull all transactions in the submarket from the last 18 months automatically. Adjustments computed against real time-on-market and concession data.
- Lease comp analysis: replace the active-listing-only rent comps with a stabilized rent table from in-place leases at comparable properties.
- Submarket forecast:replace user-input rent growth with CoStar's submarket-level 5-year forecast.
- Supply analysis: automated count of units under construction within 1 mile, lease-up timing, projected absorption impact on rent growth and exit cap.
- Sponsor track record: pull historical transactions on similar assets by the broker / seller to inform negotiation strategy and likely-clearing price.
- Underwriting sensitivity: cross-validate exit cap assumption against forward CoStar transaction cap rates in the same submarket — surfaces stale exit-cap inputs before they reach committee.
Why this matters
The current product produces a defensible memo from publicly accessible data. At firm scale, the same underwriting logic with CoStar's data layer would produce a memo that holds up against Eastdil, JLL Capital Markets, or CBRE's professional-quality output — sourced automatically rather than via the manual research pipeline that comp-and-paste workflows require.
The architecture is built to absorb CoStar without rewriting the underwriting engine. Activating it requires a CoStar enterprise contract and API credentials; the rest of the pipeline (snapshot, prose validation, PDF render) stays unchanged.