The challenge
A Product That Had Outgrown Itself
Wibey had grown feature by feature, and it showed. Screens were dense and hard to read, Chat and Sessions lived on separate pages with no shared space for a team's work, and creating anything meant leaving the product entirely, for Studio, for a separate Catalog platform, with no way to collaborate. This wasn't a fix for one broken flow, it was about making Wibey genuinely good to use.
Who are we designing for?
Target Audience
Designed for every employee across Walmart Global Tech, engineers, designers, product managers, people managers, and leadership alike.
Getting started with design
Wireframes
Every solution below went through an earlier version before it shipped. These are the ones worth showing.
Solutions
Every solution below came out of the same rhythm: a stakeholder discussion, then one day to build a working version in code, using vibe coding, so the next review could react to something real instead of a slide. That speed came with real friction. Everyone worked in their own branch, in their own style, and merging separately-built explorations into one coherent product, once stakeholders approved a direction, was its own technical problem. We also ran through our AI credits faster than expected partway through and had to bring in additional agents to keep the pace up. Underneath all of it sat a quieter constraint: several parts of Wibey belonged to other teams, so what could be redesigned outright versus what had to be re-surfaced without disturbing ownership was a negotiation on almost every page.
Overall Project Impact
Reduced tool sprawl.
Consolidating Studio and Catalog into Wibey means one less set of platforms to maintain and support.
Lower support burden.
Better cross-team reuse.
Workspaces and an interactive Catalog reduce duplicated work across teams.
Time saved at scale.
Small reductions in friction compound significantly across 10,000+ daily users.
Learnings
Speed and coordination pull in opposite directions. Vibe coding let us turn a stakeholder discussion into a working version in a single day, but everyone building in their own branch meant merging approved work back into one product was its own real technical problem, one I didn't fully anticipate going in.
Ownership shapes design as much as usability does. Several pages belonged to other teams before we ever touched them, so knowing what to redesign outright versus re-surface carefully was a constraint I had to navigate on nearly every page, not just Metrics.











