What a difference a year makes. In the first half of 2022, rents were growing by 10% or more in over 75% of U.S. markets. The picture for 2023, by comparison, has looked decidedly different. Vacancies continue to rise as rents are dropping. The culprit? A glut of new units hitting the market.

Unsurprisingly, many developers who bet big on multifamily are struggling with lease-up. Meanwhile, construction and operating costs are rising as borrowing costs soar. It's no exaggeration to say a stalled lease-up for a major multifamily building could put a developer's whole development pipeline in jeopardy, and make lenders and capital partners nervous

Rapid oscillations in the rental market are nothing new. This downturn, however, is different because it has coincided with the rise of artificial intelligence research applications. Emerging AI technologies can replace the manual processes, human biases, and misalignment of interests associated with leasing aimed at maximizing income from the asset.

The Real-Time Decision Gap

Most build-to-rent multifamily developers partner with fee managers who set the lease terms. During lease-up, developers aim to maximize profitability, while fee managers are often paid a fixed price per lease with targets for occupancy. With unilateral authority to set discounts on lease inventory, fee managers will often structure discounts to best meet their targetsand could ignore other factors like tiered rental demand, demographics, migration andrental affordability.

It's not just that fee manager compensation structures often lead to misaligned interests. The process itself is surprisingly low tech. Fee managers typically determine concession offers ad-hoc , an outdated process that fails to dynamically incorporate granular demand data and update forecasts in real time when determining optimal market rent. As a result, concessions aren't modulated as frequently as necessary to hit occupancy targets. In addition, lease terms are static. This causes 'downstream' operational issues with leases renewing in the same period 1 year laterbut also impacts customer experience. Renters want flexibility and in an oversupplied market offering shorter or longer terms may be a win-win.

As any developer also knows, with rising costs and local counties pushing back on high-rise apartments, low-density construction is increasingly the only way deals get done. While low density construction is not new, managing it well requires a muscle that needs to be developed quickly

This "old-school" approach is increasingly being threatened by AI-based machine learning (ML) applications shaking up the real estate world. A lot of hype in the AI space is just that: hype. But AI-powered real-time lease pricing is today's fact, not tomorrow's science fiction. In a sink-or-swim environment like lease-up in today's market, developers should recognize innovation as the only alternative to the status quo.

Machine learning is the co-pilot you need

Multifamily developers struggling with lease-up are in a perfect position to become early adopters of a new generation of AI platforms that employ machine learning models to calculate the most efficient trade-off between concession size and lease-up velocity.

These platforms interpret demand and iterate on achievable prices in real-time, drawing from datasets of millions of multifamily rentals, as well as metadata around neighborhood, renter affordability, demographics, and travel patterns. Multivariate forecasting and a sophisticated optimization engine allow new properties to "learn" from other areas and neighborhoods with similar characteristics. This is particularly important in information-poor scenarios, such as, build to rent, low-density or garden-style properties where very few competitors' data is available, or for lease-up assets where prior demand has not been established.

Creating a tailored lease-up is the perfect application of this technology. For a large multifamily property, this might begin with examining the price elasticity of the building's rental units and creating expected conversion curves for one-, two-, and three-bedroom apartments at different prices. The best part: The accuracy of the ML model continues to improve as users feed it additional data.

Using these demand forecasts, stochastic elasticity measurement , and other factors such as seasonality, developers can produce an optimal lease-up roadmap for their new multifamily property. Rather than being structured around the fee manager's incentives, this approach makes developer profitability the sole objective.

Triaging rents while planning

Market feasibility studies have been conducted by lenders, capital partners and developers looking to validate their assumptions. These are valuable exercises but are often costly, time-consuming and static. Questioning a market feasibility study often requires another study which could take 2-4 weeks, and cost thousands of dollars.

With the advent of AI, trawling through millions of apartments and build-to-rent homes and applying sophisticated techniques of natural language processing, machines can understand the impact of size, layout to rent. This relationship is non-linear (renters don't go looking to pay $2.0/square foot per home, they want an apartment or home with 2 or 4 bedrooms and a yard for example). Machines have the unique ability to understand this, and importantly, humble enough to confess where they are wrong on not confident about their guess. Using this, developers can triage rents real time to understand the optimal layout for maximizing demand for new projects and ultimately profit.

Retaining Residents

The benefits of AI for build-to-rent developers aren't limited to lease-up. AI-powered behavioral analytics can also help with resident retention.

Residents constantly create data through their engagement, messages, work orders, and payment patterns. AI models trained on that data can uncover trends in renter behavior so that when it comes time to renegotiate leases, management teams can predict which residents are likely to renew their leases. This information helps leasing agents direct their time and effort.

Similarly, by connecting the model to resident demographics, behaviors, and internal property data, property owners can measure the impact of a property management team's retention and loyalty efforts and predict the best initiatives to pursue to increase retention.

The Power of AI

There's no question that conditions are far from ideal for developers attempting to lease up their build-to-rent projects. However, advances in AI will allow innovative companies to confront an uncertain economic climate with accurate and actionable data tailored to the unique challenges of today's real estate developers. AI-powered lease pricing and renewal optimization software can generate significant NOI growth for developers at this critical phase of the development cycle.