The most useful thing I learned building a company wasn’t how to make the product. It was learning to notice, months in, that the product I’d built wasn’t the one worth selling.
After 25 years in technology and entrepreneurship, I thought I knew the sequence: find a problem, build the thing, sell the thing. What happened with Ruley was stranger, and the pattern underneath it now shapes how I look at almost any problem.
Ruley started small and specific. My co-founder Tracy Thomas and I were weekend pickleball players, and one too many matches had stopped dead while four people argued about a rule nobody could settle. Players disagreed, someone searched online, someone else remembered a different version, and the game sat there waiting.
That was it. We built an app so a player could ask a sports-rules question in plain language and get an answer before the next serve. No grand mission — just an irritation we happened to be standing inside. The common version of this story ends there: founder finds problem, founder builds app, roll credits. The useful part is what happened after the first problem was solved.
Start with the argument you cannot ignore
We understood the problem because we were in it, and that matters more than it sounds. There is a real difference between sitting in a room asking, “What problems could AI solve in sports?” and standing on a court thinking, “Why does four minutes of this game disappear every time nobody knows the rule?” I’ve watched founders start from the technology instead — AI is hot, so find somewhere to put AI. Backward.
Start with the irritation, and keep it narrow. Ours was: players needed a fast, neutral way to ask a rules question without digging through a rulebook. In 2025 the Sports & Fitness Industry Association described Ruley as an AI-driven rules companion that translates referee terminology into plain language so a player can ask during a game. That was the whole ambition. The game stopped; we wanted it moving again.
Solve the specific problem before expanding it
Pickleball forced us to be specific. What is the kitchen rule? Who serves next? Was that legal? Those are better product questions than “How can AI improve sports?” — and once the system worked for one rule environment, we could test whether the same problem existed elsewhere.
It did. Tennis had complicated scenarios, padel had questions involving walls, golf had an entire universe of unusual situations. Ruley’s engine now spans pickleball, padel, tennis, golf, FlingGolf, soccer and more. That expansion could look like the strategy. It wasn’t. The strategy was noticing the same problem repeating in different environments — different sport, different vocabulary, same human behavior: “What happens now?”
The product we built was not really the app
This is the realization I’d hand another founder first.
We built Ruley as a consumer experience: ask the question, get the rule, keep playing. The app proved something essential — people really will ask complicated sports questions in normal language instead of searching through a PDF. But an app creates its own problem. You have to convince somebody to download it, then remember they have it, then open it at the exact moment the situation happens. Meanwhile, leagues already have apps, federations already have digital products, clubs already talk to their members. Why should Ruley require any of them to leave what they’re already using?
That became the reframe. The app didn’t have to be the destination; the rules intelligence could sit underneath somebody else’s. We started repositioning Ruley around an API a developer can send a plain-language question to and build into their own product. Same engine, different job: a player wants an answer, a club wants fewer disputes, a coach wants a teaching tool, a federation wants its rules where its participants already spend their time.
I’ll be honest about the tense, because founders love to narrate this as a triumph. It’s a pivot in progress, and whether the infrastructure bet pays off commercially is still being written. But when the market says “actually, we want the thing under that,” it can feel like somebody misunderstood your product. Sometimes they understood it better than you did. Don’t confuse the first interface with the actual value.
Let recognition confirm the problem, not create it
Validation eventually arrived from outside pickleball: an SFIA Start-Up Challenge finalist spot in 2025, a co-sponsorship of World League FlingGolf’s LFC-5 championship that aired on ESPN2 that August. Useful signals — but none of them created the value. The court did. Recognition should confirm that a problem travels; it shouldn’t be how you convince yourself the problem exists.
And be disciplined about your own claims. We report internal accuracy above 98 percent — but that’s a company-reported figure from our own testing, not an independent audit, and I wouldn’t hand it to you as validation, because it isn’t. Know what you know, know what you claim, and know what somebody else has actually verified. Those three should never blur together just because the marketing reads better that way.
The blueprint has almost nothing to do with sports
Here’s the part that outlives pickleball. Find the repeated friction. Solve one real version of it without inflating it into a mission. Watch what people do with the solution and identify which part they actually value. Separate the engine from the wrapper. Then expand only when the same problem shows up somewhere else on its own. One caution belongs on any blueprint drawn from a single company, including this one: it’s a hypothesis from one still-unfolding story, not a law, and plenty of founders followed similar-sounding steps and hit a wall anyway.
With that caveat, the pattern travels — a court, a warehouse, a classroom, or some ossified administrative process everybody tolerates because “that’s how we do it.” The interesting companies often begin with something that looks too small: an argument, a delay, a repetitive question, a person doing something manually for the hundredth time. Small problems are not automatically small businesses; sometimes they’re just large problems viewed from very close up.
So the next time something frustrates you for the fifth time this week, don’t reach past it for a workaround. Look closer, and ask the question we should have asked sooner than we did: is this annoying me, or is it annoying everyone, every time, in exactly the same way? If it’s the second one, you’re not standing in front of an inconvenience. You’re standing on top of a business nobody has noticed yet — because friction people have stopped seeing is still friction someone would pay to remove.
Cassandra Toroian is a sports-tech entrepreneur and CEO/co-founder of Ruley, the AI “e-referee” serving tennis, pickleball, padel, golf, and soccer. With 25+ years building companies—and a background in finance (MBA) plus Python training—she’s also co-founder of Volleybird and author of Don’t Buy the Bull. A former Division I tennis player, she’s focused on using AI to make sport fairer and more accessible.