High home prices and elevated interest rates have kept Aaron Shoemaker from adding to his 30-unit co-living real estate portfolio in the Atlanta area, so back in 2024 he launched his own property management company.

Then artificial intelligence came along.

“We’ve gone crazy with AI and automation in the past year,” Shoemaker said. “I’m on Claude all day and we’ve really been able to increase our productivity.”

His company, CPM Partners, manages some 900 rooms for the co-living giant Padsplit, which has about 30,000 rooms to let in some 5,000 properties around the country. Co-living management is complicated because of the level of detail required to monitor collections and the property conditions, such as tenant disputes, parking and garbage areas.

Most recently, he has been building one app for his team to manage instead of using five different systems.

“I’m doing a lot of it myself using Claude and it’s amazing,” he said. “The two modules I've rolled out have increased our productivity by 3x.”

As an example, Shoemaker explained that he is creating an app for collections, which is an issue in the co-living space because up to 10% of renters are in arrears at any given time. (“It’s a population that really struggles to pay on time,” he said.)

His collections agent previously would have to message the tenants on several different systems, while taking the time to familiarize with the tenant’s payment history.

“The AI drafts the messaging based on the human input – this person is really trying, or thjis person is not paying,” Shoemaker said. “The AI can do the voice replication of the collections agent.”

His collections agent is currently reaching out to late-paying tenants twice a week, but with his new app, he believes those touch points can be made daily. He does have an AI voice agent in place for maintenance calls.

As for his property management business, “if we built it from the ground up now, we could run it with 3 people instead of 15,” he said. “Instead of being a supervisor of people you are managing 10 different AI agents.”

Prop.text first connected with Shoemaker back in August of 2024, when he was looking to grow his real estate portfolio, after serving as a military intelligence officer in Afghanistan, attending the Wharton School to study for an MBA and then a stint as a consultant at Boston Consulting Group.

Price points in the Atlanta region, along with interest rates that have remained above the historic lows of the pandemic period, have made it difficult to find deals that pencil. Thus the pivot to property management, a business he now sees as ripe for disruption, as startup gurus like to say.

“I’m building these tech stacks. I don’t intend to sell the tech, I intend to acquire other property management companies and apply the tech stack,” Shoemaker said. “If I’m going to do a roll up play, the software I'm comfortable with is the one I'm going to use.”

The funny thing is that he admits he was never an early adopter of tech until AI came along. 

“I’m starting to see massive productivity gains in my business,” he said. “I was always like I don’t want to manage for other people but as I started to build this business I had a lot of fun with it, and then AI and automation came along.”

The problem with off-the-shelf software solutions and apps is that they are very generic and don’t really do what most managers want. Now that the tech allows everybody to become a developer, Shoemaker does not know why others are not aggressively taking advantage of it.

“One example of the internet of things we are employing is that we have Flume water sensors that monitor our water systems,” he said. 

In the past, these monitoring systems produced a lot of false positives and a deluge (yes, pun intended) of information that made it hard for a human to track, a job that AI can do without breathing hard.

The proptech world is behind the curve on how many tasks automation can take on, he believes. While the world grows more complicated and there is more data to track, humans struggle to keep up. 

“The daily watching of things — AI makes that so much easier,” Shoemaker said.

Institutional investors are ramping up their selling now that the 21st Century ROAD to Housing Act has become law, which restricts companies that own 350 or more single-family homes from adding to their portfolios. (There was a carve-out for more build-to-rent properties.) Parcl Labs reported that for-sale listings by institutions have more than doubled since the 350-home ban was discussed in early March, when there were 4,668 active listings worth about $1.7B on the market. On July 19, that had grown to 9,586 properties and $3.1B. From February through July, the total market grew 49% while institutional listings grew 132%, almost a multiple of three of the wider market.

The productivity problem in the construction industry is a subject we’ve covered before, noting that it has fallen some 30% between 1970 and 2020, while many other industries have experienced large gains in productivity. As Brian Potter has pointed out in his Construction Physics newsletter, there are limits to the economies of scale in the process of homebuilding that would increase productivity. Potter examined the possibility of decreasing the costs of the inputs that go into construction — labor and materials — and considered the ways to extract some savings. The bottom line, Potter found, is that any gains were not significant enough to bring the price of homes down in any meaningful way. 

You can earn free prop.text merch for referring investors to the newsletter

25 referrals - hat 🧢
50 referrals - tee shirt 👕
100 referrals - weekender bag 🎒

Copy & paste this link: {{rp_refer_url}}