The App That Plays a Donkey Sound When You're Being an Ass. Then Therapists Told Him What They Actually Needed.

Ty Palmer built an app that listens to you argue with your spouse and plays a donkey sound when things get too heated.

It detects 48 spoken emotions. It shows you a transcript of your conversation with highlighted sections where you crossed the line. It even monitors your text messages and stops you mid-send to say: here's why this is wrong and here's a better way to say it. Couples therapists started using it as homework assignments.

Then they told him what they actually needed: someone to deal with the insurance companies.

That conversation happened hundreds of times. Ty is a product manager by instinct — his first move when talking to any potential customer is always "tell me the worst part of your day." And when he asked therapists, the answer wasn't tone of voice or communication patterns. It was insurance billing. Scheduling. The administrative load that was slowly grinding mental health clinicians out of the industry they'd trained for.

Donkey Chats couldn't fix that. But Ty thought something could.

That's the origin story of Clary AI — an AI admin platform for mental health clinics, born out of an app about being an ass, pivoted by listening to the people who loved the product most.

Convincing His Wife One Small Bet at a Time

Ty grew up in a catering company. Summers and weekends, cutting fruit and making sandwiches with his parents — that was just what life looked like. He never stopped wanting to build things. He just didn't know what yet.

His wife was entrepreneur-averse. She'd watched him try things. She was skeptical. So in 2019, instead of asking for a leap of faith, he proposed something much smaller: let me start a VR rental company for corporate events. I won't buy any equipment first. I'll sell the service, and if it works, I'll buy headsets with the revenue.

He made about a thousand dollars. Used it to buy headsets. Sold another event. The client needed eight headsets — so he bought more. It paid for a couple of vacations. His wife started to come around.

That's the pattern. Not a dramatic all-in moment. Just a series of small bets that kept working, each one buying a little more permission to try the next one.

He's since bootstrapped a translation company to $2 million in revenue with ten employees in its first two years (he was the first employee and eventually left), run as an independent contractor tutor, and co-founded Donkey Chats with Nathan Gunther — who, in a small world moment, had worked at Stu Kent back in the day.

Getting to the Last Interview Round on Shark Tank — Then Pivoting Anyway

Donkey Chats got far. Not just "we submitted an application" far. They made it to the last interview round on Shark Tank. The product was real, the app worked, they had users paying, and the concept was funny enough to make good television.

Then a mentor named Corbin Church gave them a different option. He said: come build the technology that mental health clinics actually need. Ty said no. Then Church said: come build the whole platform. Ty said no again. Then Church came back with what was essentially an acquisition offer — the only path forward was to roll Donkey Chats into something bigger.

They agreed, but on one condition: they weren't going to be a dev shop. They'd build out the vision they'd already developed from 18 months of listening to therapists. Donkey Chats became the foundation for Clary AI. Nathan Gunther came with it. The insight they'd gathered from hundreds of clinicians became the product roadmap.

Donkey Chats still runs. Still profitable. About ten new users a day, email sequences doing their job, paying for itself. But the founders have moved on to a bigger problem.

The $10M Raise That Closed in Days

Clary AI is now raising. Ty and Stu talked through the specifics live — Ty had just closed a $500,000 round within days of opening it, bringing total raised to $1.2 million. Angels only. Valuation: $10 million.

Stu walked him through the math that most founders gloss over.

When you raise a million at a $10 million valuation, you give up about 10% of the business. That sounds clean. But it moves the goalpost. If you'd raised at a $5 million valuation instead, you'd give up 20% — but your investors are satisfied with a $30 million exit because that's a 6x return. At a $10 million valuation, your investors need the business to grow much larger before they're happy with an exit. And "much larger" usually means "much longer."

Stu knows this from experience. He raised three small seed rounds for Stu Kent — $800K, then $1.2M, then $1.8M — all at valuations under $2 million. When a private equity group eventually bought in, they called it a bootstrapped business, not a funded one. The returns for every early investor were excellent because the entry valuations were low and the exit multiple was real.

The lesson for Ty, and for anyone raising right now: know your exit number before you set your valuation. The goalpost you plant today is the one you have to run to for the next seven years.

Why the Warehouse Manager Outsold the President's Club Rep 4 to 1

Ty's about to hire his next salesperson. He has one on the team already, self-taught in B2B SaaS, doing reasonably well. He was asking Stu whether to hire for domain expertise or raw hustle.

Stu's answer came with a story.

He hired two sales reps in the same period for Stu Kent. The first was a Macmillan guy — publisher's President's Club, experienced, had a book of business, knew how to sell. The second was a former target warehouse manager who'd done a little summer sales in college, gotten bored in supply chain, and taken a shot.

The warehouse manager outsold the Macmillan guy 4 to 1.

The reason, as Stu sees it: previous sales success often just means you were selling an easy product to a warm market with a strong brand behind you. Take away the brand and the existing book of business, and you find out what someone's actually made of. The warehouse manager had no crutch. He just had to grind.

The other piece of Stu's advice: never hire one sales rep. Always hire two. With one, you can't isolate variables. If deals aren't closing, is it the product, the market, or the rep? You don't know. With two, if one's crushing it, you know the business works. If neither is, you know something bigger needs fixing.

The Dating Game and the FU Whiteboard

Stu also shared his version of a sales hiring test that's worth stealing.

When evaluating candidates for Stu Kent, he'd give them a script, a school email address, and access to a public directory of marketing professors. He told them: I'm going to be in Portland. Oregon and Oregon State have professors there. Reach out and book me appointments. If you do, I'll come with you.

No pitch perfect required. No prior relationship. Just: here's the task, go do the thing, see what happens.

It's a fast gut-check. Either they do it or they don't. Either they figure out how to get a response or they freeze. Most great sales hires reveal themselves quickly in that environment. Most bad ones also reveal themselves quickly.

And on follow-up: Stu's old sales team had "F.U." written on the whiteboard. Standing for Follow Up. The number of deals that close on the first conversation is basically zero. The number that close after five to eight touches is where you actually make your money. A CRM isn't optional. It's how you keep track of who needs an F.U. and when.

What Stuck With Me

This episode isn't really a standard founder story. It's more like two founders talking — Stu sharing what he's learned the hard way, Ty genuinely using the conversation to think through decisions he's in the middle of making right now.

The Donkey Chats origin is funny, but it's also a real product lesson. They built something customers loved. They asked those same customers what else they hated. And the answer pointed to a much bigger opportunity. Most founders ask customers what they want built. Ty asked what they hated. Different question. Very different answer.

The raise closing in days isn't luck. It's what happens when you've already got a track record of bootstrapping successfully, when you've built trust with an early circle of angels, and when the problem you're solving is big enough that people want in before the window closes.

He's early. Clary AI is still proving itself. But Ty Palmer has done this enough times to know what it feels like when something has a real shot.

And Donkey Chats — still alive, still profitable, still catching tone problems at .87 on the criticism scale — is the best proof that he knows how to listen to people until the real pain finally comes out.

Recorded live at a Startups with Stu retreat. More at startupswithstu.com.