Field NotesDesigning Smarter
Multifamily Housing with AI

Multifamily housing has always been a balancing act. Architects working in this space know how quickly a project can stall, whether on paper or in the public process. Between restrictive zoning, rising costs, and neighborhood resistance, the system often seems structured to slow housing down rather than support its creation.

At the same time, the need continues to grow. Across the country, we are facing a housing shortage that affects renters, young people, and working families. Zoning frameworks still reflect outdated assumptions about single-family living. Density is limited to narrow corridors and often prohibited in surrounding areas. That disconnect makes it harder to create equitable and sustainable communities.

AI may not be able to rewrite zoning codes or reimagine housing typologies. But it can help us understand the systems we work within. That understanding can change what is possible. This is the intention behind Atria. It is designed to give architects better visibility into constraints, incentives, and opportunities, early enough to shape design outcomes.

I recently worked on a feasibility study for a multifamily project in Marin County. The site was zoned for 18 units, and the developer was preparing to move forward with that number. At first glance, the project seemed straightforward.

But deeper research revealed something different. By designating two units as Very Low Income (VLI), the project became eligible for a density bonus that increased the unit count to 24. This was not a workaround. It was a policy already in place, although difficult to discover without connecting multiple layers of code and regulation.

Using Atria, I was able to surface those relationships and see how the bonus applied. It pulled zoning overlays, threshold language, and relevant density provisions together in one place. That clarity made it easier to have a meaningful conversation with the client and evaluate the impact of adding six more units, not just in terms of yield but in affordability and long-term value.

I also spent time understanding the context beyond the code. Atria supported a broader demographic and economic analysis that included household income, family size, renter-to-owner ratios, nearby amenities, and transit access. That data shaped how we positioned the project.

We could now frame the proposal not just as a housing development, but as a response to actual neighborhood conditions. It aligned with local needs. It offered amenities that were missing. And it delivered rents that the surrounding community could sustain.

That narrative mattered. It gave the developer a stronger case. It made the review process smoother. And it increased the likelihood of long-term occupancy and neighborhood stability. It was a reminder that research is not separate from design. It is part of how we shape it.