Multifamily developer JPI has cut project delivery timelines by roughly 30% and halved design time through a coordinated push into construction technology and artificial intelligence. The effort directly targets the cost barriers that make workforce and affordable housing projects economically unviable at lower rent levels.
"The structural challenge of affordable housing is cost. If you can't build more efficiently, the economics don't work at lower rent levels," said Mollie Fadule, chief financial and investment officer at JPI. "Our goal is to continue to drive more efficiency, which will translate into lower cost and therefore the ability to build more affordable/workforce housing units."
JPI's first pilot project compressed the full development and construction lifecycle from about 40 months to roughly 29 months. Four factors drove the reduction: a standardized design library that cut design time per project in half, automated planning workflows built from 3D Revit models, Takt-based construction sequencing, and an Operations Support Center that provides real-time tracking and issue prevention across active sites.
AI's role in scheduling and production planning
One of the company's strongest AI applications sits inside project scheduling. Traditional multifamily schedules contain 3,000 to 4,000 activities. JPI's Takt-based production model demands far greater granularity, often 15,000 to 20,000 line items.
"Historically, building a schedule of that complexity took roughly three full-time employees working four weeks. By leveraging AI and machine learning, we've cut that effort to about one person over three to four days," Fadule said. The time saved shifts staff from manual schedule assembly toward validating and optimizing the plan.
Once construction starts, AI analyzes thousands of daily production updates from the field to identify trends, measure production velocity, and forecast schedule performance. Teams get earlier visibility into bottlenecks and can adjust sequencing and resources before delays compound.
Field data collection and the next phase of automation
JPI deploys drones, 360-degree cameras, and augmented reality glasses as core data collection tools on job sites. These feeds support daily production validation, earlier detection of installation deviations, and tighter adherence to daily schedules. The next step on the roadmap pairs drone captures with model geometry so AI can automatically flag deviations.
The company has also rolled out the Anthropic stack enterprise-wide and built internal knowledge bases and agents modeled on its operating model. A pilot implementation of Claude identified use cases that generated more than 2,000 hours of time savings per month. JPI evaluates all technology investments against improvements in schedule, cost, quality, and risk as it expands AI adoption.
People drive the train
Fadule described a model where technology lays the tracks and people drive the train. AI supplies better information and faster insights, but decisions around quality, safety, trade coordination, and field execution stay with superintendents, project managers, and the Operations Support Center.
One unexpected benefit has been subject matter experts building their own tools. "People with extensive construction experience, who may have never written code, can now use AI to build workflows, automations, and tools that solve real operational problems," Fadule said. "Because these solutions come from the people closest to the work, they're more practical, adopted faster, and help capture institutional knowledge that has traditionally lived only in individuals' heads."
AI use extends beyond the field into every business unit. Procurement teams apply it to extract and structure content from contracts. Underwriting and comp processes are seeing partial automation. The roadmap points toward continuous expansion as measurable returns accumulate.
Why this matters for real estate and construction professionals
JPI's results put hard numbers behind what many firms are still piloting: a 30% schedule compression and a 50% reduction in design time are not marginal gains. For developers and general contractors watching labor shortages and rising material costs, the scheduling application alone - collapsing four weeks of work into four days - changes how teams allocate their most experienced planners. The field data play, where drone imagery gets checked against model geometry automatically, signals where site supervision is headed. Professionals who understand how to integrate these workflows, or who pursue structured learning like AI for Real Estate Courses, will be positioned to lead implementations rather than react to them.
Your membership also unlocks:
