As the AI landscape continues to evolve, affordable housing stakeholders are exploring how the technology can address some of their challenges—from improving resident engagement and staff productivity to managing compliance, reporting, and portfolio performance.
Three industry leaders—Eric Caminiti, senior vice president of investment management at The Michaels Organization; Eric Tsai, director of strategic initiatives, enterprise, at Eden Housing; and Bianca Vanin, chief investment officer at Belveron Partners—discuss where they’re seeing the most impact, how they are determining AI investment, and the most practical tool.
Where are you seeing the most practical impact from AI today, and how is it helping your organization better serve residents while managing the operational and compliance demands of affordable housing?
Caminiti: Data analysis for large databases has become faster and more approachable for individuals without data science backgrounds. Portfolio operating analysis, financial performance analysis, and market analysis can be handled without adding additional roles/headcount. Additionally, capital expenditure planning has become simpler with tools like Tailorbird, which can create as-built plans, construction takeoffs, and replacement schedules without requiring additional labor hours from facilities and asset management.
Tsai: We’re seeing the most practical impact in three places: resident communication, staff knowledge access, and knowledge work.
We are using EliseAI’s AI assistant to manage resident communications, triage maintenance requests, and send rent due reminders. EliseAI’s multilingual capability is especially valuable for our communities, where many residents may be more comfortable communicating in languages other than English. Our residents now have 24/7 support access in their preferred language. EliseAI has reduced routine administrative tasks and given staff more capacity for higher-touch resident support.
We built a company brain in Notion that allows employees to utilize Notion AI to quickly find, retrieve and summarize policies, procedures, benefits information, property information, and institutional knowledge. The practical benefit is less time searching across scattered systems and more consistent answers for staff, which has translated into faster onboarding, better operational decisions, and more consistent service to residents.
Several departments rely heavily on Excel, including accounting, asset management, and real estate development. They have been using Shortcut, an AI plugin for Excel, to help staff analyze spreadsheets, clean up data, write formulas, summarize large datasets, and automate repetitive tasks without needing advanced Excel expertise. In practice, it has helped novice Excel users operate more like intermediates, and helped intermediate users perform more like experts without requiring years of traditional Excel training.
Vanin: The impact of AI spans multiple parts of our organization. On the asset management side, it has sharpened our time and inbox management, freeing our team to focus on higher-value strategic work, a benefit that flows through from the corporate level down to residents. For example, we’ve automated a significant portion of our internal and external reporting, cutting what used to take hours down to a fraction of that time, without sacrificing accuracy.
On the property management side, we’ve deployed tools that improve communication with residents, giving our teams real-time visibility into open issues and enabling faster resolution. This applies across the resident lifecycle, from upfront leasing and renewals to day-to-day work order management. That responsiveness translates directly into a better resident experience; less friction, higher satisfaction, and stronger retention over time.
In short, AI is helping us do more with the data we already have, aggregating it more efficiently, analyzing it more deeply, and turning it into actionable outcomes.
How are you determining whether an AI tool is worth the investment?
Tsai: We examine the following questions:
Does it reduce high-volume administrative burden?
Can we prove value through pilot metrics and staff feedback?
Is there a clear return on cost?
Is it easy enough for staff to actually use?
The tool is worth the investment if it saves enough staff time or expand our capacity in ways we may not have thought possible before. We also look for evidence that it can be adopted without adding more complexity than it removes.
Vanin: We weigh several factors, both qualitative and quantitative. The first test is adoption: whether the tool is still being used once the initial “honeymoon” period wears off, and whether that usage holds steady over time. From there, we evaluate cost against a broader set of factors: for example, time saved, integration costs and friction, error rates, and the learning curve required to get the team up to speed. Because the AI landscape and available tools are evolving so quickly, this evaluation can’t be a one-time exercise, it has to be ongoing.
What’s one AI application that has moved from pilot to everyday use at your organization, and what lessons did you learn during implementation?
Caminiti: EliseAI has moved from pilot to everyday use. AI agents are answering phones, helping with converting applications to leases, and following up on delinquent rent.
Tsai: One example is AskEden, our internal AI chatbot powered by Notion. We piloted it with a core group of users who had varying skill sets, roles, and comfort levels with technology. That helped us understand how different staff would interact with the tool, what kinds of questions they would ask, where the answers were strong, and where our underlying knowledge base needed improvement. Since then, AskEden has evolved from a tool for finding information into a tool that also helps staff create information like drafting new policies, summarizing notes, and turning existing knowledge into job aids.
The biggest lesson is that an AI chatbot is only as good as the information behind it. The rollout required us to organize our content, identify trusted sources of truth, clean up outdated documents, and help staff understand where AskEden fits into their day-to-day work.
Vanin: Across the many AI applications we’ve implemented, one lesson holds true for all of them: Take the time to set it up properly and customize it to your needs. Ready-made templates and tools rarely deliver the results we’re looking for; the tool has to be tailored to our properties, residents, and workflows.
There’s also an assumption that because it’s an AI tool, the setup and customization itself can be automated or handled by the AI.
In our experience, that has not been the case; it still takes deliberate human effort upfront. The other lesson we’ve learned is to treat implementation
as iterative rather than a one-time event: Use the tool for a defined period, review the resulting data, and
use those insights to refine the strategy and improve outcomes over
time.