AI-Powered Network Diagnostic Platform
Comcast's next-generation internal diagnostic platform gives technicians the ability to diagnose and resolve network issues at the event and device level. The platform processes real-time telemetry from over 30 million connected devices, accelerating troubleshooting analysis.
Comcast technicians relied on fragmented legacy tools to diagnose network issues. Each tool covered a narrow slice of the network stack, forcing technicians to context-switch between systems for day-to-day analysis. Average diagnosis time suffered as a result, due to workflow constraints and the absence of an intuitive tool experience.
Mapping the diagnostic landscape
Conducted a full audit on legacy tools
- Identified dozens of critical pain points across the diagnostic flow.
- Catalogued 4 distinct legacy diagnostic tools - Node Analysis (NA), Access Network Analyzer (ANA), Node Timeline, US Performance
- Documented all accessibility issues that did not align with WCAG and a11y guidelines.
Distilling signal from noise
Transformed research into actionable frameworks that would guide the platform architecture.
- Partnered with product leadership to translate business objectives into design requirements.
- Prioritized 8 core diagnostic workflows
- Established the information hierarchy for real-time telemetry and diagnostic workflows.
Workflow Experience for Main User Types
Designing the information backbone
Established the structural foundation that would unify 4 legacy tools into a single coherent experience.
- Designed progressive disclosure patterns for guiding technicians through layered diagnostic workflows, surfacing context only when relevant.
- Established unified design patterns for complex data visualizations including telemetry tables, time-series charts, performance graphs, and geospatial diagnostic maps.
- Built new components for the network diagnostic platform that were absorbed into the company-wide UI kit.
Building interactive proofs
Tested high-fidelity prototypes with key stakeholders and active network technicians during weekly syncs and sprint reviews.
- Built weekly interactive prototypes in Figma
- Collaborated with the UX research team to conduct multiple rounds of usability testing
- Iterated on data visualization patterns
Embedding intelligence naturally
Worked with ML engineers to integrate generative AI capabilities that felt native to the diagnostic workflow.
- Designed conversational AI interface for diagnostics
- Created confidence scoring UI for AI recommendations
- Built feedback loops for model improvement
Conversational AI Tool w/ Generative AI Capabilities
Codifying systematic consistency
Developed a comprehensive design system that would scale across the entire diagnostic platform.
- Created 200+ component variants
- Documented accessibility standards for field conditions
- Defined light and dark mode color tokens meeting WCAG AA contrast across all field conditions
New Components (Variations, Color Tokens)
Measurable Impact
This project is still in progress, measurable outcomes coming soon...
The most significant lesson from this next gen diagnostic tool was that enterprise AI is not about replacing human judgment. It is about amplifying it. The technicians who use this platform daily are experts. The AI serves them, not the other way around.
Role - Based Network Operations Dashboard
Got a project
in mind?
TOP 3% TALENT
Vetted byYes, you can book a free intro call through the "Get in Touch" or "Start a conversation" button. We'll discuss your goals, scope, and timeline to see if we're a good fit.
I offer product design, UX strategy, interface design, and visual storytelling, covering everything from research and wireframing to high-fidelity design and front-end implementation.
I leverage design tools, web technologies, and branding strategies to build engaging and innovative digital experiences.