FOBOH — 2024
FOBOH builds an AI-powered workforce for food and beverage wholesale and distribution. It is a B2B SaaS product for suppliers in the Australian F&B wholesale ecosystem that uses AI and NLP to take in orders and help teams process them.
My Role
Senior Product Designer, Growth & AI.
Team
1x Senior Product Designer (Myself)
1x CPO
1x Senior Frontend Engineer
2x AI & ML Engineers
2x Business Analysts
Impact
AI Order Ingestion went live with our first brewery customer in October 2024. On day one the team cleared 41 orders before the 1pm cut-off. In 11 weeks, over 1,850 orders worth $1.7M went through it. One user cleared around 171 orders in a single week and said it would have been very hard without FOBOH. Their lead called it the only successful integration they’d had.
The new look and feel left FOBOH cleaner and more consistent across the product, and the team’s verdict was that I made the application look stunning. With FOBOH Core rebuilt, engineers could build faster, and what shipped stayed close to what we designed.
How I Shaped The Product
I started with an end-to-end design audit in my first week and used it to plan the rest of the year. I gave FOBOH a new look with updated fonts and branding, a new icon set and favicons, sentence case throughout, and cleaner layouts without the extra borders. I rebuilt FOBOH Core with new modals, toasts and tags, and ran accessibility checks across our components.
In July I ran our first design research sessions with customers, using interviews, preference tests and observation to test navigation, onboarding, tables and how people understood the AI. I turned early ideas into concepts and demo flows the team could put in front of customers. On the build side, I wrote dev annotations for the full supplier flow and checked design against production, feeding quick wins into UAT.
Customers sent orders to their suppliers by email, often as attached spreadsheets or PDFs. AI Order Ingestion reads those emails, pulls out the customer, products and delivery details, and matches them to what’s in FOBOH. The supplier’s team reviews each order, fixes anything flagged, like a product or customer it couldn’t find, and completes it. The goal was to turn an inbox full of orders into a queue the team could clear fast.
For the order review screen, I explored five layouts before settling on a two-pane view. The original email, matched customer and confidence score sit on the left. The editable order sits on the right. The confidence score was my idea, so reviewers could see how sure the system was before confirming an order.
1x Case Study
(Deck Available On Request)