Shai Case Study: An Introduction

Shai didn't start as a business plan. It started as a design challenge I gave myself: what happens when you design one interface for one specific person at one specific moment in their day, with none of the compromises an enterprise roadmap usually forces on you?

That question collided with three other things happening in my life at the same time — a genuine problem I have personally, a set of new AI tools I needed to learn anyway, and a chance to build something with my wife and partner, Megan. Shai is what came out of that symphony of creative thought.

Where it started

For most of my career, I've designed marketplaces inside large organizations — Zappos, Carro, Crexi, Weedmaps. Good work, real impact, but always shaped by a roadmap, a stakeholder group, and a product built to serve many different users at once. As agentic AI tools started to mature, I saw an opening: for the first time, a single designer could realistically take an idea from concept to a fully built, native iOS, product without needing a team. I wanted to test what my craft looked like without constraints.

At the same time, I was living a problem I hadn't found a good solution for. I'm one of the roughly 40% of adults who struggle with sleep, and I was genuinely unimpressed by what was on the market. Calm, the category leader, felt expensive and generic. Smaller sleep-story apps like Slumber Studios and Sleepiest felt poorly designed. All-in-one wellness apps like Headspace asked for too much commitment before proving any value. As someone who loves stories, history, travel, and well-crafted writing, I kept thinking: this could be so much better.

The third piece was more personal. My wife has a master's degree in Modern English Literature, and I had just completed study of a master's degree in Entrepreneurship, and wanted to find a project we could build together — something that used both of our crafts and, if it worked, could become a new stream of income for our us. Shai became the place where the design challenge, the personal problem, and the partnership all met.

Grounding it in real people, not just my own frustration

I didn't want to build Shai on personal conviction alone, so I paired the market research with informal user research — conversations with friends, family, colleagues, and people I met over the course of a year. I'd ask a simple question: do you struggle with falling asleep? Consistently, about 40% raised their hand, which lined up with published research on adult sleep struggles internationally.

When I asked how they cope, the answers clustered into a handful of rituals: a sleep-story or meditation app, a familiar audiobook or TV show they've seen a hundred times, reading themselves to sleep, or simply not having a strategy at all and struggling with it night after night. Almost everyone I talked to expressed real interest in something better than what they were currently doing.

That pattern — familiar, low-effort, audio-based rituals — became the foundation for who Shai is actually built for: curious, intellectually-engaged adults whose minds don't want to shut off at the end of the day, not a broad wellness audience.

The hypothesis: story + technique, not story alone

Looking across existing sleep-audio content, I noticed most of it fell into one of two camps: a short story read slowly in a soothing voice, or content built around explicit hypnosis mechanics — countdowns, affirmations, cue words, layered under synth tones or white noise. Very little combined the two.

My working hypothesis is that both pieces matter, and that the reason ties back to how human brains evolved. We're not well adapted to the sensory overload of modern life — screens, notifications, constant input — and our brains struggle to power down inside that environment. What we are adapted to is something closer to the ancestral end of the day: a fire dying down, a tribe or family settling in together, an elder telling a story while natural sounds like wind or rain naturally occurred in the background. Shai's format is a deliberate attempt to recreate that: a genuinely interesting story to occupy and satisfy a busy, curious, mind delivered with the slow cadence and light hypnotic techniques that guide that mind down into sleep.

Where the competition left an opening

A competitive review of the leading sleep and wellness apps turned up a consistent pattern: most offerings fell into one of two traps. They were priced high enough — Calm and Headspace sit near the top of the category — that trying the product felt like a real commitment before a user had any proof it would work for them, and free stories were limited and repetitive. Or, they were paywalled outright, requiring payment or commitment to pay, before someone could even evaluate whether the content suited them.

That combination looked like the biggest barrier to adoption, and it pointed directly to Shai's model: freemium, with a real library of free stories rather than a locked demo. My assumption is that the target audience — curious people with a genuine, recurring need for something like this — are more likely to subscribe once they've actually experienced the product working for them than to commit blind. It also plays to a real structural advantage: as a two-person team running on agentic tools, our cost basis is extremely low, which means we can sustain a lower price point and put more of revenue toward acquisition than a traditional venture-backed competitor can, while still working to keep CAC disciplined against LTV. As a brand and design strategy position toward a premium experience and perception, we may perhaps explore raising the price, but my initial intention is to try a low-price model in order to gain traction and speed adoption. So many hypothesis to test!

The moment it became a real idea

The shift from rumination to commitment happened while I was exploring design directions and thinking about Apple's liquid glass material. Most apps treat glass as just another tappable UI element... the latest evolution in UI design for tappable interfaces from skeuomorphism, to flat design, back to some form of skeuomorphism. I had the idea to create a design where glass was literally used as a material and a more dynamic and universal point of interaction than simply a button. What if it behaved like a crystal-ball orb or a glass plate, sitting over a persistent video background, beautifully refracting the light from the video behind it, staying with the user throughout the entire app... It could become more than a UI element — it could embody the presence of the narrator and the story itself, a constant point of contact rather than a chrome layer.

A small amount of early exploration was enough to prove the idea worked. Getting it right took much longer — close to a year of refinement, mostly constrained by the limited time I had around a full-time design leadership role — before it settled into the minimal, dark-mode-by-default, nighttime orientation the app is today.

What I learned trying to let AI design it

Before committing to my own instincts, I tested the opposite approach: could AI tools design the app for me? I tried, and the results were consistently average — competent, generic, and forgettable, because the tools were pattern-matching against everything already on the market. The same thing happened with early story drafts: without a strong creative frame to write inside, the writing came back flat, well below the bar we wanted to put out.

That told me something important about how I wanted to use these tools going forward. Genuine creative direction — the world of the app, its tone, its visual language, the shape of a good story — has to originate from a human point of view. AI is extraordinarily good at helping articulate that vision faster, visualize it, and implement it efficiently once the direction is clear. It's a poor substitute for setting the direction in the first place. Everything that followed on Shai — the design, the writing quality, the narration — improved once I stopped asking AI to originate ideas and started using it to execute a vision my wife and I were developing.

Adapting lean principles to the constraints I actually had

My MBA's entrepreneurial concentration was built around lean methodology: validate before you build. I didn't follow that sequence strictly here, and that was a deliberate tradeoff, not an oversight. I needed to build my agentic design and development skills to stay effective in my primary career, and that meant building first. With more time available now, I'm running validation, content production, and marketing testing in parallel rather than in the traditional order.

The underlying lean principle still holds, though: minimize risk by keeping costs low. At this stage, the real cost has been my time and a low-tier AI tooling bill — no outside capital, no team to carry. The open question isn't whether the idea is viable; it's whether the content meets our own bar for quality, and whether the people testing it with us are willing to pay for it.

What’s next?

In upcoming posts, I'll articulate the tasks and initiatives required to pursue testing, validation, and to generate interest from potential customers. In addition, I'll articulate more of the specific strategy and background around the name, branding & creative, design ethos, story writing approach, and marketing tests.