A multifamily operator pushed one building's rent collection from 97.6% to 99.6%. The AI did the chasing.

A recent case study: a large multifamily operator put AI on its rent collections. In a pilot, one building's collection rate went from 97.6% to 99.6%, and it hit that number about 14 days faster than before. The result was good enough that the operator rolled the same approach out across more than 100 buildings.

Two points on a percentage doesn't sound like much. On a large portfolio, getting nearly all the rent in, and getting it in two weeks sooner, is real money and real cash flow. But the headline isn't the point. The chase is. So here's how the chase actually worked.

How it works

The mechanism is plainer than the result makes it sound. A few moves, running on their own:

1. Watch every balance, every day. The system tracks who has paid and who hasn't, and how many days past due each unpaid balance is. Nobody has to pull a report or remember to look. The moment a balance goes late, the clock starts.

2. Reach out right away, on the tenant's channel. It contacts the tenant by text, email, or voice about the late payment, early, while the balance is still small and easy to clear. Speed matters here the same way it does in sales: the longer a late balance sits, the harder it is to recover.

3. Match the tone to the situation. A tenant who's late for the first time gets a different message than someone who's late every month. The system reads the history and adjusts, so a good tenant who forgot doesn't get treated like a chronic problem, and a repeat case doesn't get a gentle nudge that goes nowhere.

4. Handle the payment-plan conversation. When a tenant can't pay it all at once, the back-and-forth about a plan is exactly the kind of repetitive, sensitive exchange that eats a manager's day. The AI can carry that conversation about the bill and the options, and keep it consistent across every unit.

5. Escalate at the threshold, with the full history. When a tenant crosses a days-late line the operator sets, the system notifies the building manager and hands over everything that was said and agreed. The person steps in for the hard call already knowing the story. Eviction and the real judgment calls stay with people.

That's it. No new portal for tenants to learn. Every late balance got worked the same day, in the right tone, and nothing sat waiting on a person to get to it.

You can do this for your own workflows too

Collections is the obvious one, but it's the same moves everywhere a balance, a date, or a promise depends on someone remembering to follow up. Chasing the signed renewal before it lapses. Closing out the maintenance request that went quiet. Collecting the document a file is missing. Each is a job that quietly leaks money when it runs on memory, and each one an AI can own end to end. If collections is where yours leaks most, that's the loop to point it at first.

And it compounds. When the same record sits under all of it, every reminder, every plan, every recovered balance teaches the system what works for your tenants and your buildings. The operator that's run this for six months knows which message lands, on which channel, at which day past due. The team starting today doesn't. The record gets sharper every week it runs.

The part the winners get right

Collections is where the relationship is easiest to wreck. Push too hard on a good tenant and you lose a renewal worth far more than the late fee. The operators who get this right don't use AI to be colder or faster at being cold. They use it to hold the relationship at a scale a person can't: remembering who's reliable, who's struggling, what was promised, and bringing a human in at the moment that judgment is needed.

That's the difference between a dunning bot and a colleague. A bot sends the same threat to everyone and forgets. A colleague remembers that this tenant has paid on time for three years and is having one bad month, and it handles that differently. The first costs you a good tenant. The second keeps the building full and the rent flowing.

Where Sonzai fits

Sonzai is the relational intelligence layer for real estate. It keeps a living record of every tenant, balance, plan, and promise, and puts an AI workforce on top of it that runs those moves around the clock, on the systems you already use. You don't build any of it. You point it at the balances that slip, set where a person should step in, and it runs.

If you want to see what your own collections look like with nothing falling through, book a demo. Bring your messiest building. You'll leave with a clear picture of the outcome, on your numbers.