So here’s the thing about automation: it’s like AI trying to order a martini at a dive bar. Technically possible, wholly unappreciated, and likely to get you punched.
This is, more or less, how Arjun Kannan sees the entire real estate tech ecosystem. In a world where every proptech startup is waving around buzzwords like “24/7 AI engagement” and “instant chatbot support,” Kannan and his startup, ResiDesk, are over here quietly yelling, “Sure, but where’s the money?”
Not metaphorical money. Not “engagement” or “CSAT” or “brand perception uplift.” We’re talking about Net Operating Income. NOI. The number that real estate operators actually care about. The number that makes them renew your contract. The number that means something when you stare at a spreadsheet and ask yourself why your units aren’t leasing.
The Gospel of NOI
ResiDesk isn’t here to “impress” tenants. It’s not here to automate them into oblivion either. Instead, it whispers sweet nothings into the operational void: Every resident message is an opportunity.
Resident wants to renew? Good. Resident wants to leave? Even better — now we know why. Resident says, “The water smells like shrimp again”? That’s a service ticket, yes, but it’s also a leading indicator that someone else in that building is one maintenance delay away from churning.
ResiDesk turns these messages — the ignored, the trivial, the 11:42 p.m. texts about parking permits — into revenue.
Not by sending emojis.
Not by being friendly.
By being… correct.
From BlackRock to Broken Locks
Before founding ResiDesk, Arjun Kannan spent his days making AI tools at BlackRock that literally helped manage trillions of dollars. He won BlackRock’s Aladdin Hackathon three times, which is like winning your company’s Olympics — if the Olympics had judges with MBAs and access to every portfolio on Earth.
Then he built machine learning systems as CTO and CPO at Climb Credit, the largest lender for bootcamp students in the U.S., where he designed an underwriting system that could predict someone’s future income. (No big deal.) The thing processed $100 million in loans and helped over 20,000 students, many of whom wouldn’t have qualified under legacy models.
Now? He’s building AI to deal with whether someone named Carol at 5B is mad about the garbage pickup. It’s quite the journey — from “revolutionizing risk analytics” to “helping property managers not get yelled at.”
But in a way, it makes sense. At BlackRock, he built systems that managed money. At ResiDesk, he builds systems that manage the things that manage the money.
Resident Sentiment Is The New Alpha
The real estate industry has historically treated residents like a rounding error. If they paid rent, great. If they left, well, that’s what the churn rate’s for.
ResiDesk flips that logic. Every resident message is a dataset. Every message is a sentiment score. Every delay is a reason someone might churn. And every good interaction is a chance to sell something new.
Where other startups throw AI at the inbox like it’s an efficiency play, ResiDesk treats the inbox like a Bloomberg Terminal. And instead of trades, they’re tracking retention, upsell, and repair triage.
(Yes, you’re allowed to feel uncomfortable about that. That’s the point.)
NOI or GTFO
ResiDesk’s value prop isn’t subtle. It goes something like this:
“Real operators care about NOI, not automation. Bots reply. ResiDesk turns resident messages into money. From leasing and renewals to move-out, ResiDesk monetizes every interaction. Automation impresses rookies. NOI impresses owners. Go beyond automation.”
This is… a lot. It’s basically an entire brand identity that doubles as a LinkedIn subtweet. But it works.
Because here’s the thing: every chatbot in this space talks about “delighting residents.” Nobody talks about collections. Or lease renewals. Or which texts predicted a 3-month-out churn risk. ResiDesk does.
And it’s already doing this across more than 160,000 units, reaching over 2.3 million residents, with clients including some of the largest operators in the U.S. and Canada. Revenue? North of $1 million ARR, growing 650% in 16 months【15†source】.
The software looks like a chat app. But underneath is a probabilistic model designed to surface ROI — not replies. And if the average property manager doesn’t notice the difference? That’s by design. Because the metrics do.
A CTO With a Product Obsession
Arjun Kannan isn’t just the CTO. He’s the guy who designed the product, built the first prototypes, debugged the message routing system at 2 a.m., and still reads chat transcripts because “that’s where the truth lives.”
He’s also managed to convince leading proptech investors like StelliFi VC to bet $3+ million on his vision — which, let’s be clear, is a vision that involves reading people’s texts about faucets. That’s not just a founder. That’s a technologist with edge.
He’s been featured in Tech Times, The Village Voice, Law360, and HackerNoon, judged panels for OnDeck and Courier, and mentored startups at TechStars — all while building a team that operates at the brutal intersection of AI engineering, compliance, and maintenance ticketing.
Which means that yes, somewhere in a Brooklyn coffee shop, a guy who was helping manage trillions of dollars at BlackRock is now debugging message sentiment classifiers about why Karen in Apt 210 thinks her ceiling is “haunted.”
That’s the world we live in.
The Bottom Line
ResiDesk isn’t building bots.
ResiDesk isn’t “doing AI.”
ResiDesk is building software that makes people more money by talking to residents better.
And if that sounds quaint, remember: when everyone else is chasing hallucinations and chatbot avatars, the guy making NOI go up tends to win.
Because in this economy? Replies are free.Retention isn’t.
To see how this all connects to Arjun’s broader journey — from Aladdin to angry tenants — check out his work at ResiDesk or his interview on Leverage with Greg Kamradt, where they discuss real-world AI, operator pain points, and why GPT isn’t a magic wand. Oh, and Greg? He’s not just a podcast host. He’s also the President of the ARC Prize Foundation, which exists to benchmark AGI-level intelligence. So yeah, they get into the weeds.
