A skincare boutique where every floor question ran through one person

The product assortment changed almost daily. New arrivals, reformulations, things that sold out and needed a substitute, and all of it lived in a team chat where each update sank under the next one. An associate who'd been off for a week couldn't tell what was current.

Worse, every judgment call ran through one person: the lead esthetician. Is this serum safe for someone pregnant? What do I recommend for reactive skin when we're out of the usual? Which regimen pairs with this cleanser? Associates left customers standing on the floor to go find her, and when she was off, the whole floor slowed down. The owner had already started cobbling together her own workaround just to keep up with the questions, a clear sign the need was real.

We built an AI floor assistant grounded on one living source of truth for the boutique's products: ingredients and sensitivities, who each product suits, what to recommend when something is out of stock, and where it sits on the shelf. When the assortment changes, one person updates the source instead of a chat thread, and every associate sees the current answer instantly.

Instead of interrupting the expert, an associate asks the assistant from a tablet on the floor: "fragrance-free option for sensitive skin?", "is this okay during pregnancy?", "we're out of X, what's the closest swap?" We paired it with simple open and close checklists so any associate can run a shift end to end, with each step logged.

The floor team recovers roughly 70 hours a month that used to go to hunting down the expert and waiting on an answer, and the lead esthetician gets her shifts back instead of fielding the same questions all day. New associates are useful on the floor after a morning rather than weeks of shadowing.

Because customers get a confident answer on the spot, including a sensible substitute when something is sold out, fewer leave empty-handed.