UKG · 2026 · Web · Design Ops
Design Velocity: Architecting a Scalable Design Ops Engine
UKG's UX team was the bottleneck on every roadmap. We were pulled in too late, sized by gut feel, blamed for slipping dates. I built the operating model and the sizing tool that moved us upstream, cut rework in half, and made 'design readiness' a number a PM could defend to their VP.
The problem
Every roadmap conversation ended the same way: 'we need more designers.' However, the data didn't support that. Designers were drowning in rework, scoping debates, and last-minute pulls. The real question wasn't how many designers we needed. It was why every project arrived at design's door already broken. That's what I was asked to answer.
Approach
- 01
Governance strategy
Defined a repeatable intake and engagement model so design entered the conversation at discovery, not at handoff. The hard part wasn't the framework, but getting Product to agree that 'is this ready for design?' should have a yes/no answer instead of a debate.
- 02
Internal Tooling (Development)
Built a logic-based sizing tool in Replit because no enterprise tool did what we needed.
- 03
Change Management
Spent as much time on stakeholder diplomacy as on the system itself. Framed the tool as PM autonomy, not a UX gate. Ran a 4-squad pilot before opening it up, so the rollout was led by PMs vouching for it, not me selling it.
Solution
Outcomes
Design rework cut ~50% in the first month
Technical review time dropped from 60 to 30 minutes per intake, by replacing subjective 'is this ready?' debates with the sizing tool's readiness score
100% adoption across the 4 pilot squads inside one quarter, with 3 additional squads requesting onboarding before the formal rollout
PMs began self-sizing complexity and bringing scoped briefs to design. This resulted in the shortening of the discovery-to-spec loop without adding headcount
Key learnings
You can't hire your way out of a broken intake process
Headcount was the obvious ask. The actual fix was making 'is this ready for design?' a question with a yes/no answer.
A number ends an argument a meeting can't
The sizing tool worked because it gave PMs something to defend to their VP when it came to planning.
Adoption is a political problem, not a product one
The framework was the easy part. The real work was making sure the first PM to use it felt faster.
Next project
Astro: From Napkin Sketch to Enterprise Platform