← Return back home
A

Bringing to life an AI-enabled marketplace that connects independent grocers with small, local brands.

Details

Role:
Lead Product Designer
Timeline:
Summer 2025 · 8 weeks
Skills:
UX, Design Systems, Product Thinking
Team:
4 — designer, 2 engineers, PM

Overview

Atlas is an AI-native B2B marketplace where independent grocers discover small brands worth stocking. My work spanned three areas of the product — onboarding, trust scoring, and shareable brand profiles — from research through a shipped beta.

Tldr

Atlas helps independent grocers find local brands worth stocking — and helps those brands get found.

The grocery middle is broken: discovery still happens over cold emails, trade shows, and luck. Atlas turns that into a marketplace with real signal.

I led design across onboarding, a trust-scoring system, and shareable brand profiles — shipping a beta to 40 grocers in eight weeks.

Atlas
Problem

Grocers and small brands want each other. The way they find each other is the problem.

Buyers told us they spend hours each week vetting brands they can’t verify, and still get burned on reliability. Brands, meanwhile, can’t get a foot in the door without a warm intro.

The opportunity wasn’t another directory — it was a layer of trust that made cold discovery safe.

Research

We ran twelve interviews with grocery buyers and eight with brand founders, then mapped the end-to-end sourcing journey. Two insights reframed the work: buyers trust peers far more than ratings, and the riskiest moment is the first reorder, not the first order.

journey map
Outcome

A beta with 40 grocers, and a trust score buyers actually cited in decisions.

In post-beta interviews, 7 of 10 buyers referenced the trust score unprompted when explaining a stocking decision — the signal we set out to build.

If I did it again, I’d test the reorder flow earlier; we over-invested in first-order onboarding before we understood where trust actually broke.