Blackbaud | SKY UX
Leadership Influence: The Director of UX Research reached out to me after seeing the impact of my audit methods. We met to explore adopting my research approach as a standard for new product development across teams.
The Director of UX Research recognized my outsized impact and invited me to scale my audit-driven approach across Blackbaud.
Mentored junior UX designers to develop product thinking and actionable research skills aligned with business outcomes.
Led enterprise design system (SKY UX), design operations, and AI integration.
Tackled fragmented product experiences from rapid M&A and legacy SKY UX patterns, which forced engineering into unsustainable custom code.
Solution: Drove company-wide audits and embedded scalable, AI-powered patterns and developer guardrails into SKY UX.
Results: Reduced custom overrides by ~40%, accelerated feature cycles, and unified AI workflows. These improvements shortened delivery times, increased product consistency, and reduced onboarding friction across teams. Early user feedback reflected increased satisfaction with smoother, more coherent workflows, while internal analytics showed measurable gains in release velocity and fewer support tickets, directly supporting key business KPIs around customer retention and operational efficiency.
Lesson: System adoption thrives when it evolves with real user and business needs, not just mandates.
Role
Role
Design System Lead & AI Pipeline Architect
Design System Lead & AI Pipeline Architect
Year
Year
2025
2025
Type
Type
Design System
Design System
Techstack
Techstack
Figma, Claude, VS Code, Storybook
Figma, Claude, VS Code, Storybook
30%
Redution in CSS Selectors
↓85%
Less components
85%
Less tech debut
30%
Redution in CSS Selectors
↓85%
Less components
85%
Less tech debut
30%
Redution in CSS Selectors
↓85%
Less components
85%
Less tech debut
Background
Why are we here?
SKY UX is Blackbaud’s mandated design system for products serving non-profits, education, and healthcare.
M&A created a patchwork of platforms; system mandates outpaced audits and product needs.
Outdated patterns blocked new workflows and AI features, pushing teams to take on technical debt through custom workarounds.
Acquired platforms lacked a unified feel.
Lesson: Mandated systems that don’t evolve can quickly become obstacles to innovation.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Background
Why are we here?
SKY UX is Blackbaud’s mandated design system for products serving non-profits, education, and healthcare.
M&A created a patchwork of platforms; system mandates outpaced audits and product needs.
Outdated patterns blocked new workflows and AI features, pushing teams to take on technical debt through custom workarounds.
Acquired platforms lacked a unified feel.
Lesson: Mandated systems that don’t evolve can quickly become obstacles to innovation.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Background
Why are we here?
SKY UX is Blackbaud’s mandated design system for products serving non-profits, education, and healthcare.
M&A created a patchwork of platforms; system mandates outpaced audits and product needs.
Outdated patterns blocked new workflows and AI features, pushing teams to take on technical debt through custom workarounds.
Acquired platforms lacked a unified feel.
Lesson: Mandated systems that don’t evolve can quickly become obstacles to innovation.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Challenge
What are we trying to solve?
The strict mandate created an operational bottleneck across engineering and design teams:
Outdated Foundations: SKY UX relied on deprecated interaction patterns that didn't support complex, modern enterprise workflows or emerging AI capabilities.
The Override Trap: Because SKY UX lacked components for newer product requirements, teams frequently wrote custom code workarounds.
Breaking Builds: Custom overrides were incompatible with SKY UX's core Angular codebase. Whenever the design system updated, product builds broke, creating massive technical and design debt.
Platform Disconnect: Newly acquired platforms felt stitched together rather than part of a cohesive ecosystem.
Lesson: Bottlenecks reveal where systems must adapt to support evolving workflows, not just enforce compliance.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Challenge
What are we trying to solve?
The strict mandate created an operational bottleneck across engineering and design teams:
Outdated Foundations: SKY UX relied on deprecated interaction patterns that didn't support complex, modern enterprise workflows or emerging AI capabilities.
The Override Trap: Because SKY UX lacked components for newer product requirements, teams frequently wrote custom code workarounds.
Breaking Builds: Custom overrides were incompatible with SKY UX's core Angular codebase. Whenever the design system updated, product builds broke, creating massive technical and design debt.
Platform Disconnect: Newly acquired platforms felt stitched together rather than part of a cohesive ecosystem.
Lesson: Bottlenecks reveal where systems must adapt to support evolving workflows, not just enforce compliance.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Challenge
What are we trying to solve?
The strict mandate created an operational bottleneck across engineering and design teams:
Outdated Foundations: SKY UX relied on deprecated interaction patterns that didn't support complex, modern enterprise workflows or emerging AI capabilities.
The Override Trap: Because SKY UX lacked components for newer product requirements, teams frequently wrote custom code workarounds.
Breaking Builds: Custom overrides were incompatible with SKY UX's core Angular codebase. Whenever the design system updated, product builds broke, creating massive technical and design debt.
Platform Disconnect: Newly acquired platforms felt stitched together rather than part of a cohesive ecosystem.
Lesson: Bottlenecks reveal where systems must adapt to support evolving workflows, not just enforce compliance.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Research
How do we know it’s a problem?
Led a comprehensive audit across 8 product teams to quantify breakdowns.
Interviewed 15+ engineers and UX leads; mapped redundant patterns and custom overrides.
Developed a Design System Health Matrix and Override Heatmap—over 60% of custom components traced to missing patterns.
Mentored Project Managers in Jobs-to-Be-Done to focus on outcomes, not features.
Lesson: Data-driven audits expose root causes and build stakeholder buy-in for systemic change.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Research
How do we know it’s a problem?
Led a comprehensive audit across 8 product teams to quantify breakdowns.
Interviewed 15+ engineers and UX leads; mapped redundant patterns and custom overrides.
Developed a Design System Health Matrix and Override Heatmap—over 60% of custom components traced to missing patterns.
Mentored Project Managers in Jobs-to-Be-Done to focus on outcomes, not features.
Lesson: Data-driven audits expose root causes and build stakeholder buy-in for systemic change.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Research
How do we know it’s a problem?
Led a comprehensive audit across 8 product teams to quantify breakdowns.
Interviewed 15+ engineers and UX leads; mapped redundant patterns and custom overrides.
Developed a Design System Health Matrix and Override Heatmap—over 60% of custom components traced to missing patterns.
Mentored Project Managers in Jobs-to-Be-Done to focus on outcomes, not features.
Lesson: Data-driven audits expose root causes and build stakeholder buy-in for systemic change.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Pivot
Should we make a U Turn?
Initial demand for “more components” masked deeper architectural issues.
Audit showed static additions wouldn’t fix legacy issues or enable AI.
Pivoted to rebuilding SKY UX’s foundation with dynamic, AI-powered patterns and flexible object models.
Lesson: True modernization begins by rethinking architecture, not just expanding features.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Pivot
Should we make a U Turn?
Initial demand for “more components” masked deeper architectural issues.
Audit showed static additions wouldn’t fix legacy issues or enable AI.
Pivoted to rebuilding SKY UX’s foundation with dynamic, AI-powered patterns and flexible object models.
Lesson: True modernization begins by rethinking architecture, not just expanding features.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Pivot
Should we make a U Turn?
Initial demand for “more components” masked deeper architectural issues.
Audit showed static additions wouldn’t fix legacy issues or enable AI.
Pivoted to rebuilding SKY UX’s foundation with dynamic, AI-powered patterns and flexible object models.
Lesson: True modernization begins by rethinking architecture, not just expanding features.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Solution
Did method fix the madness?
Designed native AI patterns (dynamic assistance, natural language input, human-in-the-loop validation, context-aware panels).
Shifted to object-oriented design reflecting real user mental models.
Organized structured co-design sessions with engineering leads, using real product scenarios and open feedback to surface conflicts early, align on technical constraints, and build cross-disciplinary trust through transparency and regular updates.
Established robust guardrails, migration guides, and centralized documentation, reducing component fulfillment time by 70%.
Lesson: Cross-functional collaboration and clear documentation accelerate adoption and reduce friction.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Solution
Did method fix the madness?
Designed native AI patterns (dynamic assistance, natural language input, human-in-the-loop validation, context-aware panels).
Shifted to object-oriented design reflecting real user mental models.
Organized structured co-design sessions with engineering leads, using real product scenarios and open feedback to surface conflicts early, align on technical constraints, and build cross-disciplinary trust through transparency and regular updates.
Established robust guardrails, migration guides, and centralized documentation, reducing component fulfillment time by 70%.
Lesson: Cross-functional collaboration and clear documentation accelerate adoption and reduce friction.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Solution
Did method fix the madness?
Designed native AI patterns (dynamic assistance, natural language input, human-in-the-loop validation, context-aware panels).
Shifted to object-oriented design reflecting real user mental models.
Organized structured co-design sessions with engineering leads, using real product scenarios and open feedback to surface conflicts early, align on technical constraints, and build cross-disciplinary trust through transparency and regular updates.
Established robust guardrails, migration guides, and centralized documentation, reducing component fulfillment time by 70%.
Lesson: Cross-functional collaboration and clear documentation accelerate adoption and reduce friction.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Principles
What was our guiding light?
Prioritized adaptable patterns over rigid controls to support diverse tech stacks.
Anchored UI in real-world object models before visuals.
Ensured the system, not individual teams, adapted to user needs.
Lesson: Design system adaptability and workflow alignment are critical for lasting adoption.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Principles
What was our guiding light?
Prioritized adaptable patterns over rigid controls to support diverse tech stacks.
Anchored UI in real-world object models before visuals.
Ensured the system, not individual teams, adapted to user needs.
Lesson: Design system adaptability and workflow alignment are critical for lasting adoption.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Principles
What was our guiding light?
Prioritized adaptable patterns over rigid controls to support diverse tech stacks.
Anchored UI in real-world object models before visuals.
Ensured the system, not individual teams, adapted to user needs.
Lesson: Design system adaptability and workflow alignment are critical for lasting adoption.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Impact
Did we hit the mark?
~40% reduction in custom codebase overrides within two quarters
Faster feature deployment and reduced UI workarounds
Unified AI framework across acquired platforms
The design system team owns quarterly audits, with active participation from product and engineering leads. Feedback loops are established through regular cross-team review sessions, ensuring findings lead to actionable updates and continued alignment with evolving needs. This structured approach sustains system relevance over time. Notably, these audits also uncovered hidden usability issues and workflow obstacles, enabling faster, more targeted improvements.
Strengthened design culture and team alignment through regular rituals
Lesson: Measurable technical and cultural wins validate the value of evolving design systems.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Impact
Did we hit the mark?
~40% reduction in custom codebase overrides within two quarters
Faster feature deployment and reduced UI workarounds
Unified AI framework across acquired platforms
The design system team owns quarterly audits, with active participation from product and engineering leads. Feedback loops are established through regular cross-team review sessions, ensuring findings lead to actionable updates and continued alignment with evolving needs. This structured approach sustains system relevance over time. Notably, these audits also uncovered hidden usability issues and workflow obstacles, enabling faster, more targeted improvements.
Strengthened design culture and team alignment through regular rituals
Lesson: Measurable technical and cultural wins validate the value of evolving design systems.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Impact
Did we hit the mark?
~40% reduction in custom codebase overrides within two quarters
Faster feature deployment and reduced UI workarounds
Unified AI framework across acquired platforms
The design system team owns quarterly audits, with active participation from product and engineering leads. Feedback loops are established through regular cross-team review sessions, ensuring findings lead to actionable updates and continued alignment with evolving needs. This structured approach sustains system relevance over time. Notably, these audits also uncovered hidden usability issues and workflow obstacles, enabling faster, more targeted improvements.
Strengthened design culture and team alignment through regular rituals
Lesson: Measurable technical and cultural wins validate the value of evolving design systems.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Reflection
What would I do differently?
Executive mandates without continuous support create hidden technical debt; system trust is earned, not enforced.
Earlier engineering involvement would have eased rollout.
As AI advances, design systems must favor intent-driven patterns over static libraries.
Lesson: Continuous feedback and collaborative iteration are essential for future-ready systems.
AI Use Transparency
Leveraged LLMs to summarize interviews and synthesize qualitative research.
Used AI tools for system documentation and rapid prototyping of natural language flows.
All final decisions, research, and logic were human-led and validated.
Lesson: AI accelerates efficiency, but leadership and accountability remain human.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Reflection
What would I do differently?
Executive mandates without continuous support create hidden technical debt; system trust is earned, not enforced.
Earlier engineering involvement would have eased rollout.
As AI advances, design systems must favor intent-driven patterns over static libraries.
Lesson: Continuous feedback and collaborative iteration are essential for future-ready systems.
AI Use Transparency
Leveraged LLMs to summarize interviews and synthesize qualitative research.
Used AI tools for system documentation and rapid prototyping of natural language flows.
All final decisions, research, and logic were human-led and validated.
Lesson: AI accelerates efficiency, but leadership and accountability remain human.
Enter image caption
Enter image caption
Enter image caption
Enter image caption
Reflection
What would I do differently?
Executive mandates without continuous support create hidden technical debt; system trust is earned, not enforced.
Earlier engineering involvement would have eased rollout.
As AI advances, design systems must favor intent-driven patterns over static libraries.
Lesson: Continuous feedback and collaborative iteration are essential for future-ready systems.
AI Use Transparency
Leveraged LLMs to summarize interviews and synthesize qualitative research.
Used AI tools for system documentation and rapid prototyping of natural language flows.
All final decisions, research, and logic were human-led and validated.
Lesson: AI accelerates efficiency, but leadership and accountability remain human.
Enter image caption
Enter image caption
Enter image caption
Enter image caption