Stealth(identity)β’
Passwords get forgotten, stolen, or spoofed. Stealth(identity) replaces them with something more reliable fingerprint, face, iris, voice, and behavioural biometrics. I led the UX across the platform's three major releases, turning a system built for field security officials using it every day.
Objective
Improve the usability of the biometric SaaS platform by introducing clear task hierarchy and streamlined workflows to reduce cognitive load and improve operational efficiency.
My Ownership
Research & Discovery
Identity management software sits at the intersection of security, compliance, and usability, a particularly difficult design space where the people implementing the system are not the same people experiencing it day-to-day. I ran structured research sessions across three distinct users separately before drawing any design conclusions.
Key Research Findings
Information Architecture
Problem Definition
The technology worked. The UX didn't, thought Stealth(identity) had powerful biometric capabilities. The product's competitive positioning had shifted from government system integrators to enterprise IT buyers, and the UX had not moved with it.
Problem 1 β Admin Console: No Task Orientation
The existing console presented all data, all the time, with no hierarchy. An administrator managing 10,000 enrolled identities had no way to surface what required attention, what was pending, or what had failed. Every action required navigating multiple nested menus.
Problem 2 β Enrollment Flow: No Real-Time Feedback
Biometric capture quality is critical, a poor-quality fingerprint scan creates a weak identity that fails authentication later. Enrollment operators had no inline quality indicators during capture. They submitted captures, waited for server-side validation, received pass/fail with minimal context, and restarted. Enrollment rates were significantly below business targets.
Problem 3 β Authentication: No Trust Architecture
For the person being authenticated, the moment of biometric capture was completely opaque, no explanation of what was being captured, no indication of progress, no confirmation of success. The interface communicated nothing. Trust was absent by design.
Problem 4 β Multi-Modality Inconsistency & Unusable Audit Log
Face recognition, fingerprint, iris, voice, and behavioural biometrics each had separate UIs built by separate teams at different times, completely different mental models, terminology, and interaction patterns. Simultaneously, the compliance audit log was a flat, unsorted, unsearchable table of raw transaction data that required developer support to generate any meaningful report.
Design Process & Solution
Given the complexity of the platform and the diversity of user types, I structured the design process in distinct phases, starting with the admin console, moving through enrollment flow, then authentication UX, and finally the design system that unified everything across all modalities.
Design Principles
Old Design
Unified Design System
One of the most impactful deliverables was establishing a unified component library across all biometric modalities and platform surfaces. Before the redesign, each modality had been built independently. After, a shared set of tokens, components, and patterns ensured that moving between fingerprint and face enrollment felt identical β reducing cognitive load and training requirements significantly.
Biometric Enrollment Wizard
Simplified steps with clear progress and focus
Scanning passport for MRZ
Fingerprint capture
Portrait capture
Admin Console Redesign
The console was rebuilt from the administrator's actual workflow outward. The home dashboard now surfaces actionable status immediately, enrollments pending, authentications failed in the last 24 hours, system health, and alerts requiring action. Navigation was reorganised around tasks rather than data types.
The flat audit log was replaced with a structured compliance dashboard. Filterable by date range, modality, outcome, user, and operator with one-click export to formats required by common compliance frameworks. Compliance teams can now self-serve most reporting requirements without any developer involvement.
Outcomes
Let's build something meaningful together.
Open to discussing new opportunities, design challenges, and collaborations.
