Pareto Intelligence

Pareto Intelligence builds data products for some of the largest healthcare payers in the U.S., including Cigna, Blue Cross Blue Shield, and Humana. Its tools help these organizations reconcile millions in financial discrepancies, dig into complex claims data, and run more efficiently.

When I joined, the analytics underneath were strong, but the experience on top was a patchwork. Dashboards didn't match each other, workflows were confusing, and dense datasets made it hard for analysts to read a screen and act on it. My job was to bring order and usability to a set of enterprise tools handling huge volumes of sensitive financial and clinical data.

There was no shared design system, so teams built components on their own and the same UI element behaved differently from one product to the next. The mismatches in color, spacing, type, and interaction piled up design and engineering debt, slowed onboarding, made errors more likely in financial workflows, and pushed maintenance costs up over time.

At the same time, the platform had to handle dense healthcare data: claims, diagnoses, risk scoring, audits, payments, RAF charts, and compliance triggers. But the underlying structure was weak, so it was hard to spot anomalies, decide what to fix first, or compare trends. Everything leaned on Tableau dashboards that worked but didn't scale, with no clear hierarchy, clumsy cross-filtering, and little support for how specific people worked.

The company was moving to Looker and needed a UX lead to shape that transition. There was a bigger problem underneath it too: navigation, filtering, and insight patterns changed from tool to tool, so analysts had to keep relearning how to work as they moved between products.

I audited the portal end to end, pairing heuristic and task-time analysis with interviews and shadowing across analysts, claims auditors, actuarial teams, and compliance partners. The same issues surfaced everywhere: insights weren't prioritized, filtering was inconsistent, tables were overbuilt, too many clicks stood between analysts and the detail they needed, and visual noise buried the metrics that mattered.

From there, I restructured the information around a simple flow: insight, then context, then action, then audit. I standardized filtering, simplified navigation, and added reusable data groupings and priority-based layouts for the metrics that mattered most. I validated it with interactive Figma prototypes tested with analysts across teams.

I built a single design system that every product adopted. It standardized the type hierarchy; color coding for statuses, risk states, and data confidence; and reusable dashboard pieces like cards, KPIs, comparison tables, and filters. It also set navigation patterns, spacing, grids, composition rules, and interaction behaviors like hover, expand, drilldown, sort, and compare.

It cleared our worst internal bottlenecks: quicker, clearer specs for engineering, consistent layouts for analysts, and more predictable timelines for leadership on new features. It was easier to maintain too, and it became the foundation for every product update over the next two years.

I redesigned several of the key dashboards so scattered data turned into something people could act on. They didn't just look better. They were measurably faster to use, easier to read, and closer to how analysts actually think, which led to sharper decisions.

I led the vision for moving dozens of Tableau dashboards into Looker: rebuilding the visuals with consistent logic, clearer drill paths, shared charting rules, better comparison views, and stronger performance under load.

Two things made this harder than a visual refresh. Legacy data didn't map cleanly from Tableau to Looker, so reaching parity took careful reconciliation before anything could be made better. And analysts fluent in the old dashboards had to adopt new patterns, so the work had to earn trust, not just ship. The design system and prototype testing carried much of that weight, keeping the transition legible and giving people something concrete to react to before it went live.

Measured in usability testing and product analytics, the redesign cut task-completion time by 25% and error rates by 30%. New subscriptions rose 20% over the two years that followed; design was one contributor there, alongside sales, pricing, and product changes, not the sole cause.

Just as important operationally, the design system gave engineering a single, consistent foundation for every release that came after.

The lasting lesson wasn't about any one dashboard. It was that in enterprise analytics the interface is only as trustworthy as the system beneath it, and a design system is what makes that trust repeatable across products, teams, and years.

If I ran it again, I'd bring analysts into the migration earlier. The redesign landed, but adoption is its own design problem, and the people living in the old dashboards needed more runway to trust the new ones.