Alcy has taste

When we introduced Alcy, we started with a specific shopping problem: you know the item you want, but finding it in your size at a price you are comfortable paying takes hours of investigation.

Our new curation expands where that conversation can begin. You can come to Alcy with a reference, an idea or a question about what to buy. Alcy can help you explore brands and understand the differences between your options.

Building this capability raised a question we care about deeply: where should Alcy’s taste come from?

Consider a request for lightwash blue jeans like the Japanese brand Auralee.

An agent could interpret that as a search for a similar cut and color. It might return a pair from Old Navy or Gap. Those could be perfectly good jeans. But the recommendation may have missed much of what you meant by “like Auralee.”

Perhaps you care about the fabric, the way it falls, the treatment of the wash or the care behind how the garment is made. You might be looking for another Japanese label with a similar approach to materials and construction. A photograph and a color match only capture part of that.

For Alcy to give useful advice, it needs to understand which qualities make the reference meaningful to you.

We are building around small data

We are building a knowledge store that gives Alcy a deliberate foundation for these conversations.

We think of it as small data: a selected body of knowledge about brands, designers, materials, construction and the relationships between them. Tastemakers help shape this pool by defining the references worth including and the distinctions worth preserving.

Two brands might share a silhouette but approach fabric very differently. Another pair might look different at first glance while sharing a design lineage or an approach to making clothes. Those relationships become much more useful when we retain the reasons behind them.

The selection matters. So does the context. A recommendation should be able to explain why a brand belongs in the conversation, where it differs from your reference and what you would be trading off by choosing it.

More data can dilute the point of view

Frontier models bring an extraordinary breadth of knowledge. But a broad knowledge base does not, by itself, give a shopping assistant a coherent point of view.

Our concern is what happens when the details that matter to a particular shopper are overwhelmed by the references that appear most often. A widely distributed brand has many opportunities to show up across the web. A small label making a particular kind of Japanese denim may have far fewer.

The frequency of a reference tells us little about whether it belongs in your wardrobe.

We want Alcy’s advice to preserve those smaller distinctions. That means making intentional choices about the knowledge it consults and the qualities it considers when comparing brands. The knowledge store gives us a way to make those choices explicit.

Your taste gives Alcy direction

The pool we are building gives Alcy a foundation. Your preferences give the conversation direction.

You might love a brand’s proportions and dislike its fabrics. You might want something more unusual, something easier to wear or something that works with pieces you already own. Those details should change the advice.

We started Alcy to help people reach the stores and products they would otherwise miss. The concierge extends that idea to the decision itself: helping you discover what else is worth considering, with enough context to understand why.