If you’ve been following me for a while then you may be aware that I spent the first years of my career as an environmental engineer, working as a government contractor. If you’ve ever sold anything to a government agency, you already know that the technology you’re selling matters far less than whether the person across the table trusts you enough to bet their reputation on you.
That’s exactly the pattern I recognized the moment Arjun Kannan told me how he landed his first customers at ResiDesk.
In fact ResiDesk has been built on early traction, largely through paying customers. That’s a rare sentence in a vertical that’s as hostile as the real estate industry is to new vendors.
Real estate operators get pitched constantly. Most of them have been burned by a “revolutionary” platform before. Most B2B SaaS founders struggle to crack this industry at all, because the usual playbook (SEO, paid ads, a slick demo) reaches everyone except the people who need to actually buy what you’re selling.
Arjun and his co-founders cracked it anyway. Here’s what they figured out, in the order they figured it out.
The Pitch That Landed Skipped the Technology Story
Arjun’s answer to “how did you land your first customers” had two parts.
- The boring part → They called as many people as they could and kept dialing until someone took a bet on them. Many successful companies start this way.
- The interesting part → They led with a problem every landlord already recognized, before ever even mentioning that they leverage AI.
Arjun’s team defined that problem statement: “Your renter lives in your building for 365 days a year, and you only talk to them on day one and day 365. Everything in between goes dark, until they’re angry enough to complain.”
Every landlord Arjun talked to had felt this, regardless of how much software they’d used and that’s the point… He was describing their own Tuesday back to them, instead of teaching them a new category.
Steal this. If you’re thinking of selling a solution into a legacy industry, always test whether your buyer already lives the problem you’re thinking of solving before you test whether they understand your proposed solution. The pitch that converts is usually the one that names their pain in language they were already using.
Peace of Mind Is What Real Estate Actually Buys
Landlords buy peace of mind, low change risk, and the freedom to stay hands off. Most property managers are already running their business across 15+ browser tabs. Asking them to learn one more dashboard isn’t what they want to hear.
So ResiDesk made a call early on that Arjun credits as one of their most important. They built entirely inside of the channels people already use.
- Property managers get reached through email, a channel already open on their desktop all day.
- Residents get reached through text, a channel already living in their pocket.
- The product lives inside existing habits, instead of creating a new one to learn.
That single decision did more for their close rate than any feature on the product roadmap ever could.
Steal this. Ask what channel your buyer already trusts and uses constantly. Then build the thing that shows up there, instead of asking them to come to you.
Is Your Startup Falling Behind Its AI-Native Competitors?
The One Thing Every Founder Selling Into This Space Gets Wrong
Real estate is what Arjun calls an offline social network. Most property managers and owners maintain a thin, almost invisible web presence, and SEO and paid ads mostly miss them entirely.
But they know each other. Every landlord in a given zip code or building type is bound together by the same regulations, the same regional associations, the same conferences.
Arjun described discovering this by accident.
- He’d talk to Client A, easy to find, with a real website.
- Client A would mention Client B, who existed only through word of mouth, invisible to search.
- Client B knew Client C, equally invisible online.
- A, B, and C all knew each other. Only A had a website.
90% of the market you’re trying to reach might be sitting one degree away from the 10% you can actually find, and the way it runs is through relationships as a primary GTM motion, before you ever consider ad spend.
The correction he’d make if he could go back a year came down to two moves.
- Skip the giant trade show booth.
- Show up at the small regional gatherings instead.
At a conference with 150 booths, everybody forgets who they talked to within the hour. At a room of 20 or 30 regional operators, you can build a relationship that turns into a customer eight months later.
Steal this. If your buyers live offline, meet them offline. Find the smallest room where they already gather, and show up as a person (not a booth vendor).
The Technical Unlock That Mattered More for Sales Than for Engineering
ResiDesk’s real technical breakthrough was defining the common property-management layer first, then using LLMs to map messy inputs into it. It was quite surprising to me to learn that, while this breakthrough was significant to product engineering, its commercial impact also showed up most clearly in sales. The use of LLMs in this way removed a sales objection.
Before LLMs, connecting to a fragmented industry’s messy systems meant becoming a certified vendor for every property management platform and hoping their APIs exposed the whole picture. Most didn’t. Property managers trust their own reports over their own APIs, because reports are what they actually look at.
Once Arjun’s team could define the handful of things every property manager needs to track, the list turned out to be short.
- units
- residents
- rent
- tickets
They used AI to map any spreadsheet, database, or report into that structure. That let them say something only a rare few in their category could say, “we’ll work with whatever you already have”… which in turn means that rigid upfront integration is much less of a blocker.
And that single claim is what got skeptical buyers to actually try them.
Steal this. The AI capability that matters most in your GTM is probably the one that removes the objection that stands between “interested” and “signed.”
The Stat That Only Means Something Because of the Mechanism Behind It
In measured deployments, ResiDesk has seen 60% higher resident satisfaction, 6x more positive reviews, and 5x fewer maintenance round trips. Those numbers are even more impressive once you understand where they come from.
Before ResiDesk, here’s what happened.
- A resident’s AC breaks.
- They mention it at the front desk.
- Someone gets scheduled, shows up at the wrong time, misdiagnoses the problem.
- A second appointment gets scheduled for the actual fix.
- A part is missing, so a third appointment gets scheduled.
- Two to three round trips is normal, and the resident gets angrier at every step, mostly because everyone stays silent about what’s happening.
With ResiDesk, here’s what happens instead.
- The resident texts about the AC the way they’d text a friend.
- A short back-and-forth (too hot, too cold, making noise) produces a real diagnosis before anyone shows up.
- The technician is more likely to arrive with the right context, parts and equipment, at a coordinated, convenient time.
- The resident gets a heads-up the whole way through.
The mechanism behind all three stats is the same. Round trips drop, because the problem gets clearly understood the first time. Fewer round trips means a faster fix. A faster, clearly communicated fix means a happier resident. A happier resident is far more likely to leave a review, and dramatically more likely to leave a five-star one.
Arjun’s read on why this works is that satisfaction has less to do with the absence of problems (residents stay pretty forgiving when something breaks) and everything to do with staying in the loop while it’s happening.
Steal this. Performance metrics only earn the label “proof” once they trace back to a specific, explainable mechanism. Until then, they’re just marketing claims.
Why Traditional Cash-for-Intro Referrals Didn’t Convert, and What Worked Instead
I asked Arjun if ResiDesk runs a standard referral program, the kind where you pay a customer a fee for sending you a lead. His answer was simple. They tried it more than once, and each attempt fell flat.
The reason, in his own words, was direct. “It’s seen as transactional. People see it coming from a mile away.” He kept coming back to the fact that real estate is a relationships business. A cash-for-referral exchange breaks the exact trust the product is built to create.
What worked instead was treating their most engaged customers as something closer to investors in the product by providing:
- Reference calls
- Real case studies
- Better long-term contracts, as an acknowledgment of the relationship, ahead of a transaction for it
He drew a useful distinction here between two different types of people.
- People using the product day to day (property managers, on-site staff) respond far better to a case study ask or a reference call than to a gift card.
- People making the buy decision (owners, portfolio managers) are the ones where a more standard B2B incentive can still work.
Two different audiences, but two totally different currencies.
Steal this. Before you build a referral program, ask whether your best customers want more money or want more recognition. The wrong choice here is why most referral programs in relationship-driven industries don’t work.
When the Founders Stopped Being the Entire Sales Team
Arjun pinpointed the exact customer count where personal involvement in every sales call became optional. That number was 30 customers. That’s when the company felt they’d hit what he calls “level one” of product-market fit. A salesperson other than a founder can explain the value and get a yes.
Past that point, the motion split cleanly across four roles.
- Outbound goes through a BDR and SDR team.
- In-person and conference relationships go through account executives.
- Upsell goes through account managers.
- Founders still join calls today, but for a different reason.
These days they join for what Arjun calls “level two conversations”, his term for helping a customer see patterns across all their resident data. It works a lot the way a CRO learns to read trends across a sales org, rather than simply checking whether one rep uses the tool.
The Advice Arjun Would Give Any Founder Selling AI Into an Old, Trust-Driven Industry
I asked him to close with a playbook, one thing to double down on and one thing to ignore.
Here’s what to double down on.
- The parts of the relationship that stay human, regardless of how far automation goes.
- The human, trust-driven edge of your category is the part competitors chasing pure efficiency will always undervalue.
Here’s what to ignore.
- The instinct to sell efficiency alone.
- Faster and cheaper is the most common AI pitch in every legacy industry right now, and Arjun’s read on it is direct. “If all AI solutions are competing on is doing things faster and cheaper, you’re in a race to the bottom.”
- Anyone with access to the same models can match you on speed. Only a rare few can match you on trust.
His framing, which I loved, was this. Efficiency and effectiveness live in entirely different categories.
- Answering a resident’s question in two seconds is efficient.
- Using those two extra minutes you just bought to actually understand what that resident wants, and feeding it back into decisions about renewals, amenities, and retention, is effective.
The efficiency race is a race to zero. The effectiveness race is the one that must be won.
The efficiency race is a race to zero. The effectiveness race is the one that must be won.
The Pattern Underneath All of This
Every lesson in this piece traces back to the same idea. ResiDesk out-related the real estate industry, instead of trying to out-automate it with AI speed and agility.
The technology let them do three things at scale that used to require an army of people:
- Read messy data.
- Text every resident.
- Catch problems before they become complaints.
The reason any of that mattered commercially is that it let them rebuild something the industry had slowly lost… It restored the sense that somebody with two residents (and a personal relationship with each of them) runs a better business than somebody with two thousand residents and a dashboard.
If you’re building AI for an old industry, Arjun framed this as a question for you to consider: Are you making the transactions faster, or are you making the relationships achievable again at a scale that used to simply be beyond reach? Arjun’s bet is that only one of those is defensible.
P.S. Near the end of our call, Arjun mentioned they’d just made their first marketing hire, and the job now is turning everything that’s working in their sales calls into outward-facing messaging. He validated the story with real customers and real revenue before he spent substantial money on a marketing hire. I told him that’s exactly the right order to do it because – well, it is.
