Workforce Housing AI screening to reduce evictions is no longer a future concept in the workforce housing—it’s already transforming how leading operators manage risk and occupancy. In this case study, discover how AI-driven tenant screening helped a large workforce housing portfolio improve leasing decisions, reduce eviction risk, and unlock stronger financial performance—without limiting access for qualified renters.
See how smarter data, fair housing–aligned models, and seamless workflow integration are changing the game for multifamily, BTR, and affordable housing developers.
In the workforce housing rental market, where costs are rising and tenant outcomes directly impact portfolio performance—Pangea Properties needed a better way to screen applicants. Pangea partnered with Beekin Labs to enhance its internal
workflows with an AI-powered tenant screening add-on purpose-built to predict eviction risk at the point of lease origination. Initial results showed that Beekin Labs’ solution delivered a threefold reduction in eviction filings relative to the decline in approval rate—and generated $4.8MM in higher NOI for the portfolio.
Download
Key Results
"We were very early adopters of not just collecting data but using it to power our decisions in a very dynamic way—including flexible marketing, promotions, and pricing to adjust to supply and demand changes on a dime."
The Challenge: Screening at Scale Without Sacrificing Fairness
For workforce housing operators like Pangea, tenant screening is a continuous, high-volume process shaping thousands of leasing decisions and directly affecting portfolio performance. The challenge is 2-fold: screen out genuinely high-risk applicants while keeping qualified renters from being unnecessarily turned away. Legacy screening tools left significant performance on the table: approving tenants who would later be filed for eviction, and rejecting applicants who would have paid reliably. Any model deployed at scale must also guard against inadvertent bias against protected classes—a legal and ethical imperative that required careful, deliberate design from thoutset.
The Solution: Beekin Labs AI-Screening Add-On
Beekin Labs designed and calibrated its resident screening solution using a few thousand leases—cases that had
either matured beyond six months or resulted in an eviction filing within the first year. The model’s target: predict
whether an eviction would be filed within 12 months of tenancy. Across the sample, the baseline eviction filing rate
was 19.5%.
Before any modeling work began, Beekin removed all attributes that could introduce bias against protected classes
under the Fair Housing Act:
Geolocation attributes — city, state, zip code, zone, and property identifiers — and any amenity
descriptions that could indirectly signal location
Accessibility amenity data that could indicate a tenant’s disability status
From the remaining 104 attributes, Beekin’s team engineered an additional 4,650 nodes for training AI —
dramatically expanding the model’s predictive surface through data segmentation, expert-driven domain insights,
and features designed to generalize robustly across portfolio conditions.
Download the case study to learn more.
f we could talk to every prospect one-on-one and gauge what they're looking for, we would know exactly who's going to stay and for exactly how much. Beekin Labs turns qualitative factors into a score that allows us to evaluate every application and make business-driven moves.
Precision Cuts, Not Broad Strokes
A key value proposition of Beekin Labs’ add-on solution is delivering meaningful risk reduction without requiring
operators to broadly restrict their tenant base. The curve you can see if download the case study shows the relationship between approval rate and eviction rate: on average, each point of approval rate reduction yields three points of eviction rate reduction.
What It Means for BTR Workforce Housing Operators
Pangea workforce housing results illustrate what is possible when AI is applied to leasing workflows. A 75% increase in approvals alongside a 40% reduction in evictions is not a tradeoff — it is an optimization made possible by better information.
For any operator managing hundreds or thousands of units, reimagining leasing and operations workflows can result in customer centricity and higher Net Operating Income.
Beekin Labs’ approach — decades of machine learning, real estate and rigorous feature engineering represents a blueprint any platform can embed directly into its tech stack. As AI workflows become imperative, this is a massive
win in time to value. By leveraging spatial, operational and financial datasets, Beekin Labs brings faster time to
market to CRE technology and data companies.
About Pangea Properties Workforce Housing Operator
Pangea Properties is a 13,000-unit residential property operator. With a focus on providing quality housing at scale, Pangea applies data-driven operational practices to improve resident outcomes and portfolio performance. After successfully scaling up over 15 years, growing to 3 states (Indiana, Illinois, and Maryland), Pangea Properties workforce housing operator successfully exited the business to a larger private equity-backed competitor. The superior technology and operating platform excellence was a defensible position for Pangea workforce housing to build from.
About Beekin Labs
Beekin Labs from Beekin builds bespoke AI systems for real estate. Using Beekin’s unified data and decision platform, it helps multifamily, single-family, and industrial operators with better operations and underwriting. Beekin Labs connects to your existing systems, normalizes your data into a governed ontology, and deploys best-in-class AI models trained on millions of rows of industry data to optimize CRE asset performance. Products have generated millions in efficiency and automation across leasing, screening, and asset valuations, and power technology firms including Green Street.
Download the case study to learn more.