ISPM Dashboard
Saviynt
2024-2025
UI/UX Design Lead
16 Stakeholders
As the primary designer for Saviynt's ISPM product, I led the design of a dashboard experience that brought the full scope of the platform's identity security ecosystem into a single, actionable view. Spanning risk detection across identities, accounts, roles, entitlements, and cloud resources, it was Saviynt's first true dashboard and its first multi-tenant experience.
Delivered despite a mid-project tech stack pivot and a demanding monthly release cadence, the result was an information-rich foundation built to scale.
The Challenge
Multi-Tenant, Uncharted Territory
ISPM was Saviynt's first multi-tenant product which was a meaningful architectural departure from its core platform, which had always operated on dedicated instances per customer (single tenant). There was no internal playbook for designing at this scale.
No Dashboard Precedent
Saviynt had dashboard-adjacent elements such as simple widgets and basic visualizations, but nothing that qualified as a full dashboard UX. The vision called for an out-of-the-box (OOTB) experience first, followed by customizable dashboards powered by AI and traditional UI tools. We were building the foundation and the roadmap simultaneously.
Compressed Design Timelines
A monthly release schedule versus the platform's usual trimester cadence meant design windows sometimes collapsed to a matter of days. The approach had to be run-and-gun by necessity: design fast, iterate frequently, and always architect for scale.
Platform-Wide Scope, Lean Team
Platform-Wide Scope, Lean Team ISPM touched every identity, access, and asset security surface in the Saviynt platform, including its policy engine. I was the primary designer for the entire product offering, with occasional support from one or two others during crunch periods.
The Process
COMPETITIVE ANALYSIS & THIRD-PARTY EVALUATION
The engineering team initially selected Apache Superset as the dashboard solution, with the plan to adapt its visuals and UX to fit Saviynt's brand and user journeys. I familiarized myself with its flows for building and managing custom dashboards, while also evaluating Metronic for its visual layer compatibility and ClickUp for its data visualization patterns and drill-down workflows. Both offered useful reference points for how to approach layout and interaction at dashboard scale.
TECH STACK PIVOT
Midway through, the team moved away from Superset in favor of an in-house solution. Superset proved too rigid, mainly:
The visual customization needed to match Saviynt's design system was extensive, requiring a large development effort that we didn’t have resources for.
Its UX architecture wasn’t flexible enough for our customers' workflows.
In hindsight, the pivot was a net positive. It gave us the space to rethink the dashboard experience on our own terms rather than working around someone else's constraints.
BUILDING THE UI LIBRARY
With no existing dashboard component foundation to draw from, the library had to be built in parallel with active design work. I collaborated closely with our design system owner to develop a scalable, consistent set of components. Together, we established key patterns around widget interactions, data and trend notations, and the distinction between functional and status colors.
The Solution
By the start of Q4 2025, the ISPM Dashboard had taken shape across three interconnected pieces.
The OOTB Dashboard
The out-of-the-box experience gave customers immediate visibility into their identity security posture, surfacing risk across identities, applications, and cloud resources. Findings were paired with actionable mitigation recommendations spanning governance domains like Privileged Access Management, account abuse, and SOD conflicts. Posture Insights added another layer, delivering targeted metrics across specific cross-sections of access risk for customers who needed to go deeper.
SaviAI Chart Generation
Beyond the fixed dashboard, customers could generate their own custom charts directly from their connected data using a SaviAI-powered workflow, then add them to the dashboard as widgets. It was an early proof of our AI framework in a high-stakes, data-rich context, and a preview of where customizable dashboards were headed.
“Coming Soon” Patterns
Given the pace of feature releases, we developed a set of "Coming Soon" treatments to announce and preview upcoming capabilities within the dashboard itself. It kept the experience feeling current and forward-looking rather than incomplete; a small but important trust-building detail for early adopters.
OOTB DASHBOARD
AI-GENERATED CHARTS
Outcome & Reflections
By the end of FY 2025, the ISPM Dashboard was live and in use by 12 customers, including GE Healthcare and Cigna. Early reactions confirmed we had built something people wanted, but the feedback that followed made clear where the next iteration needed to go.
Actionable Insights
Customers responded well to the drill-down behavior, where clicking any widget opened a modal with more granular data. But visibility alone wasn't enough. They wanted to act on what they were seeing, not just understand it. Deeper action workflows had been part of the UX architecture from the start, and were actively in development at the time of my departure.
Feature Bloat
Half of our initial customers felt the dashboard surfaced too many metrics, which was a signal that we were prioritizing coverage over clarity. Rather than presenting widget after widget, the smarter path is being deliberate about what earns visibility. Quality of insight matters more than quantity. Had I continued on the project, consolidating and refining the widget set would have been the priority.
Handling Big Numbers & Trends
Our customers were large organizations like healthcare systems, enterprises, etc. where they could be tracking identities and resources up into the millions. At that scale, meaningful changes registered as fractions of a percent, making standard trend visualizations almost unreadable. I updated our number notation patterns to include upper and lower range indicators for both small and large fluctuations, and worked with the development team to scale data visualizations relative to the highest and lowest values over a 30-day window — rather than against the full resource pool. It was a subtle but significant shift in how customers could actually read and understand their data.