
Your customers can't name
what they want. So they leave.
From a confusing counter to a sales channel that didn't exist before.
Why it happens
Every category has its insider language. In fragrance, it's notes, families and accords: a vocabulary built for experts, then handed to customers. Most people can't use it. So they give up, buy what they already know, or lean on whoever is behind the counter, and many never get to the brands on the shelf.
This retailer saw it every day. Customers walked in knowing how they wanted to feel, and walked out without the words to find it.
- Where does your customer get lost in your language?
- How many sales end because someone couldn't describe what they wanted?
- What would change if they could ask in their own words?
How we designed it
We started by challenging the assumption the whole category is built on: that people choose a fragrance by what's in it. We listened to shoppers in store, and studied how people talk about scent everywhere else. Both told the same story. They don't search for ingredients. They describe a moment, and how they want to feel in it.
With AI-assisted research, we gathered the language of communities, reviews and videos, and decoded it: how people describe a mood, a memory, a presence. We mapped the shopper's journey from "I don't know what I'm looking for" to the bottle in hand, and everyone it touches: the customer, the staff behind the counter, the retailer and the brands on the shelf. We found the gap at the exact point where expert language takes over.
The design turns that gap into a conversation. Three questions about a moment, a mood and a presence lead to a scent direction from the retailer's own catalogue. We piloted fast: an AI experience built in weeks, not months, live in store behind a QR code and online, so any customer could try it without staff or an app. Behind it, we built a dashboard that turns every conversation into insight: what customers feel, ask for and choose, and how the channel is being adopted, session by session. We tested both with fragrance experts and with new customers. Both were impressed.
What changed
The pilot proved it: 227 guests, 391 sessions, 170 personalised recommendations. And the most common thing people did wasn't search for a note or a brand. It was write down a feeling.
The retailer now has a sales channel that wasn't available before. It works in store and online, guides customers who would have walked out, and, through the dashboard, shows how its customers really think and talk about scent.
For you, it may not be fragrance, or AI. It will be whatever turns your customers' own words into their next decision.
Status
Pilot completed. Next pilot with perfume brands in preparation.
Next case
A marketing agency · Singapore
Clients ask for posts.They need someone to name the problem.