Why We Built Harry George on Medusa

Harry George makes luxury essentials out of Montreal, the pieces worn closest to the skin. It started with underwear.

The founder is Harry George Owen. For years, he performed with Cirque du Soleil, on stage in front of thousands of people, night after night. And every single night, the same small problem followed him out there. His underwear did not fit the way it was supposed to. Not once, not twice, but every performance, for years. Eventually he stopped waiting for someone else to fix it, and made his own instead.

That is the whole story, and it is also the brief we were handed, even if nobody wrote it down that way.

A man who would not accept a badly made piece of clothing should not have to accept a badly made website either. He built his first product by refusing the ordinary version and making the right one instead. We tried to build his store the same way.

A fashion house lives on detail. The weight of a fabric. The finish on a seam. The way a garment sits on a person's shoulders. That kind of care does not stop at the product. A customer can feel the moment a beautifully made thing is sold to them through a clumsy website, the same way Harry could feel the moment a garment was almost right, and still wrong.

So we spent time on things most people will never consciously notice. Here is one example. Every animated element on the site, the small moving pieces of the design, is built to sense the colour of whatever section it happens to land on, and adjust its own colour to match. On most sites, a designer picks one colour for a graphic and it stays that colour everywhere, which means it looks perfect on the section it was designed for and slightly wrong everywhere else. Ours does not do that. It looks right every single time, on every section, because it is quietly checking its surroundings and adjusting itself.

Here is another. Most websites are only carefully designed for three or four specific screen widths, a phone, a tablet, a laptop, a large monitor. Anything in between those exact sizes tends to look slightly off, a little cramped or a little stretched. We built this site to scale smoothly across every width in between, so there is no in-between size where it looks wrong. Whatever device someone is holding, whatever size their browser window happens to be, it looks intentional.

And then there is the search box and the shopping bag, the two things a customer touches the most on any store. Speed alone is not the interesting part here. Everybody has come to expect a store to be fast. What almost nobody expects is for those two small things to feel like part of the same considered world as everything else on the site, rather than a plain, generic box that was simply dropped in because every store needs one. Ours were designed to match the rest of the experience, in how they move and how they feel to use, not only in how quickly they respond.

None of these things ever get written into a project brief. Nobody asks for them by name. But they are what it actually looks like when someone refuses to accept the ordinary version of something, even in a place nobody was expecting them to look closely.

Here is a simple way to picture what we built. Think of one central kitchen that can supply as many different restaurants as the brand ever decides to open. Right now, that kitchen supplies three of them.

The first restaurant is the public storefront, the one any customer can visit. It works in three different currencies and two different languages, so someone shopping from Montreal, Paris, or New York sees prices and words that make sense to them, without needing a separate website built just for their country.

The second is a private portal for wholesale buyers, the shops and retailers who buy in bulk. They get their own login, their own pricing, and their own set of rules, separate from what a regular customer sees.

The third is the admin area, where Harry's own team manages everything, the products, the stock levels, the orders, the pricing, all from one screen.

Each of these three draws from the same underlying system, rather than three separate systems that all need to be told the same thing over and over in three different places. The team still enters information the way any team does, adding a product, updating a price, adjusting stock, but they are doing it once, into one system built for this brand specifically, instead of duplicating that work across three unrelated tools that were never designed to talk to each other.

And this is really the heart of it. Harry has talked about wanting a mobile app one day. He has talked about the brand growing into new markets, and new categories beyond where it started. Because of how this was built, none of that requires starting over. A mobile app, a new market with its own currency, a new product line, each of those can simply connect to the same kitchen already in place. The foundation was built with room for the brand to keep becoming more than it is today.

This next part never shows up in a walkthrough, because nobody ever demonstrates their accounting software to a client. But it matters more than almost anything else in the build. The orders, the stock, and the customer records all talk directly to the systems Harry's team already uses to run the rest of the business, instead of being typed in twice by a person. A store that constantly fights its own back office is not really a refusal of ordinary. It is ordinary, just with nicer lighting.

Something changed while we were partway through this build. It is worth slowing down for, because when we first sat down to plan this store, it was not something anyone on our side was designing around yet.

People have started buying things without ever visiting a website. Instead, they ask an AI assistant, the same kind of tool you might use to answer a question, and the assistant does the shopping for them. It checks the price, checks whether something is in stock, and completes the purchase, all without a human ever loading a single page.

This is not a prediction about the future. It is already happening. Google launched something called the Universal Commerce Protocol at a major retail conference in early 2026, working together with Shopify, which lets these AI assistants read a store's catalogue directly, check real prices and real stock, and complete a purchase on a customer's behalf. Google is rolling out the same kind of checkout inside its own AI tool, Gemini. OpenAI has done the same thing inside ChatGPT. And over Cyber Week, one measurement found that roughly one in every five online purchases, worldwide, involved an AI assistant somewhere in the process.

Here is why that matters for a brand like Harry George. An AI assistant does not browse a website the way a person does. It does not scroll down a page. It does not look at a photograph and feel something. Instead, it asks a very direct question. Something close to, does this particular store have the product someone is looking for, is it actually in stock, is the price fair, and can the purchase be completed right now. Then it reads the answer straight from the store's own information.

If a store cannot answer that question clearly and quickly, something important happens, and it is not what you would expect. The store does not simply appear lower down in a list of results, the way it might with an ordinary search engine. It gets left out of the conversation entirely. The assistant never mentions it at all.

Here is the part worth being honest about. When we chose to build Harry George this way, keeping one central system that anything could connect to, we were not thinking about AI assistants at all. Almost nobody in our industry was yet. We built it this way for a simpler reason, because we did not want a platform deciding what this brand was allowed to become next.

It turns out that same decision is exactly what a store needs today to be read clearly by an AI assistant. Not because we predicted this moment, but because building something open enough to connect to anything, almost by accident, means it is also open enough to be understood by the new kind of shopper that has just arrived.

If you run a fashion brand, and you find yourself thinking about a build like this one day, here is something worth remembering. It comes from a conversation worth having with whoever is quoting you for the work.

Do not only ask what it will cost to build the first version of your store. Ask what happens later, on the day you want to add something that was never part of the original plan.

If the honest answer to that question involves the word rebuild, meaning they would need to tear down what they already built and start over, then what you were actually sold was a single launch. Not a business that can keep growing.

None of this means every brand needs a custom build like this one. If Harry George were only ever going to sell one focused range of products, directly to customers, with no real plans beyond running a simple store, a hosted platform like Shopify would have served him just as well, and for less money. There is nothing wrong with that choice. Most brands should make it.

A build like this one starts to make sense the moment you can name three real things you want to do next with your brand, and at least one of them makes you a little uneasy when you picture trying to do it on the platform you have today. Perhaps a second product line that ships very differently from the first. Perhaps wholesale customers who need their own pricing. Perhaps a new country, with its own currency and its own language. Perhaps, one day, an app. Once that list stops being a maybe, and becomes something you actually intend to do, the right answer changes too.

Today, Harry George runs three separate experiences from one shared system, and a catalogue that is ready for whatever new categories the brand grows into next. A mobile app has not been built yet. But when Harry is ready for one, it will be a new conversation about design and experience, not a conversation about starting the whole system over from nothing.

Harry spent years being told that the thing he wanted did not exist, until he decided to make it himself. It felt right that his store should be built to work exactly the same way.