Tenant Intelligence Engine
A classification layer over tenant data — predicts churn probability on existing tenants, scores incoming leads, recommends retention actions tied to predicted risk drivers.
- Engagement
- 6–10 week build · ongoing operation
- Built for
- Multifamily operators · Property managers · Leasing teams
Tenant lifecycle decisions — which leads to chase, which renewals to push hard, which complaining tenants are about to leave — are made on instinct or rough rules. The data that would inform them (payment history, support-ticket patterns, neighborhood signals, lease terms) is collected but not analyzed.
What this is
A tenant-lifecycle classification platform. Two outputs:
- Churn prediction. Per-tenant 60-day and 180-day non-renewal probability. Probability-banded so the operator acts on confidence levels.
- Lead scoring. Per-lead expected-lifetime-value score. Surfaces priority leads for the leasing team's attention.
Plus a recommendation layer that ties predicted risk drivers to action options (timing of renewal outreach, concession-magnitude tradeoffs, retention communication patterns).
How it's built
LightGBM with monotonic constraints (for interpretability — operators need to defend decisions), feature engineering on the operator's tenant-and-payment data, fair-housing audit of feature importance against protected-class proxies. Integration into the property-management system for surface (AppFolio, Yardi, ResMan).
What you get
- Per-tenant churn probability with banding.
- Per-lead lifetime-value scoring.
- Retention-action recommendation layer.
- Fair-housing audit documentation.
- Quarterly model refresh.
Engagement is shape, not list.
Length and price are functions of the data and the destination. The shape below is the typical engagement.
- Length
- 6–10 week build · ongoing operation
- Lead
- Bogdan
- Cadence
- Async, weekly
- Bar
- Production
Scoped during the discovery call against the actual data and the operation it integrates with.
Principal engineer. Architecture and most code ships through one keyboard.
Written updates between, calls when the decision needs the room.
Async correctness, capacity under burst, observability at every boundary.
Products this composes with.
Same suite, or vertical-specialized versions in another.
What buyers ask about this one.
What's the churn-prediction accuracy?
On the standard multifamily population, the 60-day churn prediction typically has AUC around 0.82–0.88, with the high-risk decile being meaningfully predictive of actual non-renewal. The product surfaces probability bands, not point predictions — the operator acts on bands, not single-tenant scores.
What data does it use?
Lease terms (length, current vs. market rent, concession history), payment history, support-ticket and maintenance-request patterns, tenure, demographic indicators where the property tracks them (employment stability proxies, household composition). All operator-owned data; we don't pull external personal data on tenants.
What does the lead-scoring half look like?
Incoming leads (tour requests, application submissions) get a score representing expected lifetime value — composite of likely-to-convert, likely-to-sign-long-lease, likely-to-renew. Surfaces high-value leads for prioritization by the leasing team.
How do you handle fair-housing compliance?
Carefully. The model excludes protected-class proxy variables (we audit feature importance against the standard protected-class list). The retention-action recommendations are surface-level (timing, communication tone, concession-magnitude options), not personal-treatment differences. We document the fair-housing audit per engagement.
Pricing?
Scoped to portfolio size and integration depth. Discovery call covers both.
If the deliverable matches the gap, the next step is one call.
We'll scope length and price against your data and the operation it integrates with. No retainer, no fishing.
Bogdan and team · async-first · OP—2026