Plan Management: Insurance Services
A benefits administration platform at ADP, serving 50,000+ HR practitioners across the US and Canada. I led the experience design end to end.
At most companies, someone in HR sets up the health and insurance plans everyone picks from each year. At ADP they do this for tens of thousands of companies. The catch is they are usually not insurance experts, the rules are fiddly, and a small mistake changes someone's real coverage.
I redesigned the tool these HR teams use, pulling a job that was spread across disconnected systems into one place. I focused the design on making the right choice obvious for people who do this only occasionally, and validated that direction with real internal users before committing.
- Argued for a retail-like experience over a standard admin tool
- Deferred categorisation and self-service to keep MVP shippable on time
- Built new Design System components alongside the product work
- Unified plan management platform for 50,000+ users across US and Canada
- Automated plan modification, eliminating manual placeholder creation
The business constraint was a legacy platform decommission. The design opportunity was bigger: rebuild the experience from the ground up so practitioners could manage plans and help employees make informed healthcare decisions, all in one place.
- End-to-end experience design: bringing a fragmented multi-tool workflow into a single platform
- Interaction design, flows and visual design through to final delivery
- Significant component contribution to the ADP Design System, which was at a nascent stage when this project ran
- Research planning and synthesis with internal users across a large user base
- Insurance domain SMEs: business logic behind plan structures, age-banded enrolment rules and carrier sync
- Platform engineering: owned the legacy platform decommission and migration
- PM: scope sequencing and delivery timeline decisions
- Design system governance: what shipped as reusable components
Setting up a benefit plan should feel like filling out one clear form. Instead, the people doing it were stitching the same information across two systems that did not talk to each other, then chasing sign-offs by hand.
HR practitioners were managing benefit plans across two disconnected tools, manually reconciling data and chasing approvals at every step. There was no unified view of plans, no automation, and the employee-facing experience offered no real way to compare options or understand what fit.
The job was spread across two systems that did not talk to each other. The redesign pulled it into one place with a single view of every plan.
- Started with a unification brief: bring the disconnected experience into one platform
- As we mapped use cases, the retail-vs-admin-tool question surfaced in brainstorming
- Made the case that newer agents and admins needed an experience that made the right decision legible, not just possible
- Pressure-tested this direction with internal users (a large population at ADP) before committing, so the retail-style framing was grounded in how practitioners actually decide, not just my hypothesis
- Insurance domain SMEs involved at every step for business logic accuracy
- Design System was at a nascent stage, requiring new component contributions alongside product work
- Platform engineering owned the decommission and migration running in parallel
- Self-service and recommendations layer scoped jointly with PM before being deferred on budget grounds
The brief was to unify a disconnected experience. As we worked through the use cases, a more fundamental question emerged about what kind of product we were actually building.
Reference screenshot of the legacy plan management experience that was being replaced.
Putting the self-service layer in phase two was the right call for the budget and timeline we had, and we built the foundation to take it later. What stuck with me is how little it would have needed to land. Even a contextual prompt or a recommendation flag at plan selection would have pushed the experience closer to the retail feel we were after. Next time I'd raise that earlier, and I'd frame the cost of leaving it out as plainly as the cost of building it, because that's the conversation that actually moves a roadmap.
This project was delivered in 2022. The high-fidelity screens are not available to share here. The flows and architecture above reflect the full end-to-end journey. The decision content in this case study, the fork, the tradeoff, the research that shaped the final structure, is where the real work lived.
That is the summary. The full case study covers the research, the design decisions and the final screens in detail.
Password protected. Available on request via email or LinkedIn.