AI & E-commerce/Client/2025
ChatBlu
Trust Is the Feature.
Overview
ChatBlu started as an AI platform for syncing inventory across Shopify, Amazon and everywhere else a merchant sells. On paper it's a perfect job for AI. In practice it makes merchants nervous, and they're right to be. If an automated system breaks a listing, it's their account standing and their cash flow on the line.
So the real problem wasn't syncing data. Trust was the feature. I came in as interim design lead for the brand, the product strategy and the UX/UI, and almost every decision came back to one rule. Anything the AI does has to be visible, and it has to be reversible.
Design thinking
-
Trust beats trends.
AI has a look by now (purple, sparkly, dark), and none of it says dependable to someone running a warehouse. I went the other way, with a deep-ocean world, a bright blue palette, high-contrast type, and a crab mascot whose pincers grip a bunch of data sources at once.
-
Never let the robot guess.
Onboarding makes merchants choose a source of truth for every field. When channels disagree, the UI shows the conflict and offers clear choices instead of quietly making the call for them.
-
Get the chat out of the bubble.
A chat thread is great for context and bad at the big picture. So the conversation sits on the left, the real product form sits on the right, and you watch the AI make its edits as it goes.
-
Receipts on everything.
Any field ChatBlu touches gets a little label. Tap it to see what it used to be and what else the AI suggested. Nothing is permanent unless you want it to be.
-
Simple up front, deep underneath.
The same product table has to work for someone launching their first TikTok shop and a warehouse manager mapping API fields across thousands of SKUs. Layered hierarchy and progressive disclosure keep the power tools tucked away until someone goes looking for them.
Impact
The brand and product vision helped the founders raise $500k in pre-product funding. Early users started onboarding and managing their catalogs in Q1 2026, and the company has since grown into hospitality SaaS, with hotel partners in California, New York and Europe.
I owned brand architecture, product strategy and UX/UI, grounded in jobs-to-be-done interviews and surveys with merchants. The biggest lesson I took away is that AI takes some weight off users, but it also makes them work backwards from what they want, and it costs trust the first time it's wrong. Slapping it on as an “ask anything” box hands all that thinking to the user. You still have to design the use cases.
