Back to work

Keeyu

AI agent for post-purchase ecommerce operations.

Role
Product Designer
Deliverables
End-to-end product, design system
Timeline
2026

Overview

Keeyu runs the back end of ecommerce: payments, warehouses, carriers, and the support tickets they generate. Working end-to-end across the design and product thinking on a rebuild that shifts the product from a reactive dashboard into a proactive AI agent.

Discovery

Operations teams watch multiple tools and react to issues as customers report them. Most fires are preventable — the data to spot them is already in the system, just not surfaced in time.

Goals

  • Move from reactive to proactive

    Design the surfaces where Keeyu flags risk and proposes action, instead of waiting for ops to find it.

  • Make the agent trustworthy under pressure

    Every recommendation traces back to data the team can see, override, and reason about in seconds.

Outcomes

  • Reframing the product around proactive detection across payments, warehouses, and carriers.
  • Establishing the patterns for AI recommendations + human override that the rest of the product can build on.

A note

Keeyu is an active work in progress. Specific product detail and metrics live under NDA. Happy to chat about my design thinking, process, and the decisions behind the work.

Next case study

AtriumRebuilding the design system and product flows for a fast-scaling team.