From gut feel to 100% accurate tenant screening
A Canadian property management company with 50 years in the market. We built the ML model
that replaced gut-feel tenant screening with 100% accurate risk prediction.
100%
Model Accuracy Rate
1 month
Delivery From Scoping to Handoff
1 team
Dedicated End-to-End
AI & Automation
About The Company
The Challenge
Screening at scale needs a system
At hundreds of applications across multiple properties, consistent risk evaluation can't rely on individual calls. The business needed a repeatable, data-driven process that worked the same way every time.
Stated income isn't the full picture
Credit scores and declared earnings tell you what a tenant wants you to see. Actual bank transaction data shows how they manage money month to month, which is what matters for rent reliability.
Risk signals need to surface earlier
The earlier a risk signal surfaces in the application process, the more options a property manager has. The goal was to move the decision point earlier with better data behind it.
Scaling before the model was right
In a high-competition health app market, more spend without smarter targeting just inflated CPC and diluted ROI. The goal wasn't more leads: it was the infrastructure to find the right ones, consistently, at scale.
The Approach
Data & Scope
Mapped applicant data sources, defined what "delinquency risk" looked like in the data, and scoped the ML model architecture.
- Bank transaction data access agreed
- Risk indicators defined with the client
- Model architecture scoped
Model Build
Built the ML algorithm to analyse applicants' bank transaction data, identifying spend patterns that reveal true income level and payment reliability, not just stated figures.
- Transaction pattern analysis built
- Income verification model trained
- Spend behaviour risk scoring developed
Validation & Delivery
Model tested against historical applicant data. Accuracy validated at 100% before handoff. Maintenance scope agreed.
- Backtested against historical cases
- 100% accuracy confirmed
- Handed over with documentation
The Results
100%
Model accuracy confirmed against historical applicant data before go-live.
1 month
From scoping to validated delivery: one dedicated team, start to finish.
1 model
Single algorithm replacing inconsistent manual screening across all properties.
0 guesses
Every screening decision now backed by transaction data, not intuition.
Your highest risk decision is who you let through the door.
If you're still screening tenants on paper, we should talk.
Case Studies
Results that Compound
Machine Learning For Customer Predictions
Company builds health & fitness apps used by millions. We built the ML infrastructure that made their Google Ads spend chase value.
