Epic AI Design System: Unifying 80+ Apps from Zero
Disclaimer: Due to IP restrictions, actual Epic designs and assets cannot be shared, unless they've been made public.
Epic, the largest EHR provider in the U.S., needed a way to make AI safe, consistent, and adoption-ready across 80+ clinical apps.
I built the company’s first AI design system from scratch, turning siloed AI projects into a unified, trusted foundation now used by 80K+ clinicians and 330+ healthcare organizations.
The Results
The AI design system became Epic’s foundation for all AI-driven experiences:
Trusted by 80K+ clinicians across 330+ healthcare orgs.
Rolled out before Epic’s largest annual conference, with 40,000+ attendees seeing the system in action.
Endorsed by Epic’s C-suite as a “beautiful” solution that aligned speed with adoption.
Established validated design standards that continue to guide all future AI initiatives.
"Working with Josh was a fantastic experience. He was highly responsive to feedback and played a key role in shaping our ideas into a compelling narrative and an enhanced user experience around AI.
I especially appreciated his initiative in researching the best solutions and his creativity in exploring new directions."
Nick Marzotto
Head of AI Customer Success, Epic
The Process
I used a design thinking approach with rapid prototyping, iterative feedback, and constant user validation to ensure we were building the right design system for real-world traction.
Step 1: Define the foundation
Audited 30+ in-progress AI features to uncover inconsistencies.
Identified key adoption blockers: lack of trust, unclear ownership, and no shared UX standards.
Step 2: Create reusable AI patterns
Defined 4 core AI workflows (summarization, task automation, content transformation, drafted text).
Built reusable components for explainability, trust signals, and safety checks.
Step 3: Establish visual + UX identity
Designed AI-specific icon system, visual language, and the “Bloom” symbol as the universal AI marker.
Standardized interaction rules across all apps.
Step 4: Scale + validate with users
Rolled system out across 80+ apps with dev + design teams.
Tested continuously with clinicians, iterated for safety and clarity.
Solving Hard Problems at Scale
While the actual design artifacts are confidential IP, these examples show the kinds of complex UX problems I solved, where design decisions directly impacted adoption, trust, and safety:
AI Design System
Before: Each team built AI their own way. Different layouts, controls, and voices. No shared framework. Good ideas got lost, bad ones repeated. Chaos at scale.
After: I created a unified AI design system spanning 80+ clinical apps. Teams reused components, workflows, and safeguards instead of reinventing. This gave Epic a shared language for AI and turned scattered experiments into a scalable system.
Trust & Explainability
Before: AI outputs showed up with no context. Clinicians didn’t know how reliable they were, so trust wavered.
After: I designed explainability patterns, such as confidence scores, source transparency, and clear controls (accept, adjust, dismiss). Clinicians could see why the AI suggested something and act with confidence.
Visual Identity
Before: AI had no “face.” It blended into Epic’s UI, so users couldn’t tell what was human vs. AI. Misinterpretation was a constant risk.
After: I partnered with designers to create a distinct AI identity called Bloom. Every AI output carried a consistent marker. Clinicians quickly learned to spot and interpret AI suggestions without confusion.
Scalability
Before: Every new AI feature was treated as a one-off pilot. Teams debated the same questions again and again. Progress slowed.
After: I shifted the mindset from “designing features” to “designing systems.” Built repeatable workflows, patterns, and safeguards that scaled. Teams moved faster, with trust and usability baked in from the start.
Broader Scope of Work
Beyond hands-on design, my role covered the larger ecosystem of AI at Epic:
Partnered with leadership to shape AI strategy across the company.
Built pitch decks and roadmaps used with partners (e.g. Microsoft) and major health systems.
Acted as an internal “AI design consultant” for 80+ teams and senior leadership.
Delivered presentations to R&D, Mayo Clinic, and executive audiences to educate and evangelize AI design.
Built and managed systems, like Figma component libraries, internal guidelines, and reusable design assets.
Created future vision roadmaps to define where Epic’s AI should go next.
This gave me both a seat at the strategy table and ownership of the systems that made execution possible.
Disclaimer: Due to IP restrictions, actual Epic designs and assets cannot be shared, unless they've been made public.
Disclaimer: Due to IP restrictions, actual Epic designs and assets cannot be shared, unless they've been made public.
Epic, the largest EHR provider in the U.S., needed a way to make AI safe, consistent, and adoption-ready across 80+ clinical apps.
I built the company’s first AI design system from scratch, turning siloed AI projects into a unified, trusted foundation now used by 80K+ clinicians and 330+ healthcare organizations.
The Results
The AI design system became Epic’s foundation for all AI-driven experiences:
Trusted by 80K+ clinicians across 330+ healthcare orgs.
Rolled out before Epic’s largest annual conference, with 40,000+ attendees seeing the system in action.
Endorsed by Epic’s C-suite as a “beautiful” solution that aligned speed with adoption.
Established validated design standards that continue to guide all future AI initiatives.
"Working with Josh was a fantastic experience. He was highly responsive to feedback and played a key role in shaping our ideas into a compelling narrative and an enhanced user experience around AI.
I especially appreciated his initiative in researching the best solutions and his creativity in exploring new directions."
Nick Marzotto
Head of AI Customer Success, Epic
The Process
I used a design thinking approach with rapid prototyping, iterative feedback, and constant user validation to ensure we were building the right design system for real-world traction.
Step 1: Define the foundation
Audited 30+ in-progress AI features to uncover inconsistencies.
Identified key adoption blockers: lack of trust, unclear ownership, and no shared UX standards.
Step 2: Create reusable AI patterns
Defined 4 core AI workflows (summarization, task automation, content transformation, drafted text).
Built reusable components for explainability, trust signals, and safety checks.
Step 3: Establish visual + UX identity
Designed AI-specific icon system, visual language, and the “Bloom” symbol as the universal AI marker.
Standardized interaction rules across all apps.
Step 4: Scale + validate with users
Rolled system out across 80+ apps with dev + design teams.
Tested continuously with clinicians, iterated for safety and clarity.
Outcomes and Impact
This work established a unified AI design system that reshaped how 80+ clinical applications at Epic delivered AI. Instead of scattered features, teams could now scale safely through shared components, visual identity, and built-in safeguards. The result was faster development, reduced duplication, and consistent adoption across critical healthcare workflows.
Beyond usability, the design system influenced Epic’s broader AI strategy. It anchored partner presentations with Microsoft and Mayo Clinic, positioned design as the “AI expert” voice for leadership, and laid the foundation for future vision roadmaps. The impact extended from daily clinician workflows to enterprise-level partnerships, proving that UX can drive both adoption and strategic direction at scale.
Lessons for Founders
The patterns from this project extend directly to AI startups building their first or next product:
Think in systems, not features: Designing reusable patterns and safeguards pays off, whether you’re scaling from 1 product to 10 or 10 to 100.
Build trust into the workflow: Adoption hinges on explainability and user control. Treat them as core features, not extras.
Position for impact: A strong design system can influence partnerships, customer buy-in, and the credibility of your AI strategy.
This work proved how UX can turn fragmented AI pilots into a scalable system that clinicians trust. At Epic, these design decisions shaped strategy, partnerships, and the path for how AI enters healthcare safely.
Epic AI Design System: Unifying 80+ Apps from Zero
Epic AI Design System: Unifying 80+ Apps from Zero
Company
Role
Lead UX/UI Designer
Disclaimer: Due to IP restrictions, actual Epic designs and assets cannot be shared, unless they've been made public.
Epic, the largest EHR provider in the U.S., needed a way to make AI safe, consistent, and adoption-ready across 80+ clinical apps.
I built the company’s first AI design system from scratch, turning siloed AI projects into a unified, trusted foundation now used by 80K+ clinicians and 330+ healthcare organizations.
The Results
The AI design system became Epic’s foundation for all AI-driven experiences:
Trusted by 80K+ clinicians across 330+ healthcare orgs.
Rolled out before Epic’s largest annual conference, with 40,000+ attendees seeing the system in action.
Endorsed by Epic’s C-suite as a “beautiful” solution that aligned speed with adoption.
Established validated design standards that continue to guide all future AI initiatives.
"Working with Josh was a fantastic experience. He was highly responsive to feedback and played a key role in shaping our ideas into a compelling narrative and an enhanced user experience around AI.
I especially appreciated his initiative in researching the best solutions and his creativity in exploring new directions."
Nick Marzotto
Head of AI Customer Success, Epic
The Process
I used a design thinking approach with rapid prototyping, iterative feedback, and constant user validation to ensure we were building the right design system for real-world traction.
Step 1: Define the foundation
Audited 30+ in-progress AI features to uncover inconsistencies.
Identified key adoption blockers: lack of trust, unclear ownership, and no shared UX standards.
Step 2: Create reusable AI patterns
Defined 4 core AI workflows (summarization, task automation, content transformation, drafted text).
Built reusable components for explainability, trust signals, and safety checks.
Step 3: Establish visual + UX identity
Designed AI-specific icon system, visual language, and the “Bloom” symbol as the universal AI marker.
Standardized interaction rules across all apps.
Step 4: Scale + validate with users
Rolled system out across 80+ apps with dev + design teams.
Tested continuously with clinicians, iterated for safety and clarity.
Solving Hard Problems at Scale
While the actual design artifacts are confidential IP, these examples show the kinds of complex UX problems I solved, where design decisions directly impacted adoption, trust, and safety:
AI Design System
Before: Each team built AI their own way. Different layouts, controls, and voices. No shared framework. Good ideas got lost, bad ones repeated. Chaos at scale.
After: I created a unified AI design system spanning 80+ clinical apps. Teams reused components, workflows, and safeguards instead of reinventing. This gave Epic a shared language for AI and turned scattered experiments into a scalable system.
Trust & Explainability
Before: AI outputs showed up with no context. Clinicians didn’t know how reliable they were, so trust wavered.
After: I designed explainability patterns, such as confidence scores, source transparency, and clear controls (accept, adjust, dismiss). Clinicians could see why the AI suggested something and act with confidence.
Visual Identity
Before: AI had no “face.” It blended into Epic’s UI, so users couldn’t tell what was human vs. AI. Misinterpretation was a constant risk.
After: I partnered with designers to create a distinct AI identity called Bloom. Every AI output carried a consistent marker. Clinicians quickly learned to spot and interpret AI suggestions without confusion.
Scalability
Before: Every new AI feature was treated as a one-off pilot. Teams debated the same questions again and again. Progress slowed.
After: I shifted the mindset from “designing features” to “designing systems.” Built repeatable workflows, patterns, and safeguards that scaled. Teams moved faster, with trust and usability baked in from the start.
Broader Scope of Work
Beyond hands-on design, my role covered the larger ecosystem of AI at Epic:
Partnered with leadership to shape AI strategy across the company.
Built pitch decks and roadmaps used with partners (e.g. Microsoft) and major health systems.
Acted as an internal “AI design consultant” for 80+ teams and senior leadership.
Delivered presentations to R&D, Mayo Clinic, and executive audiences to educate and evangelize AI design.
Built and managed systems, like Figma component libraries, internal guidelines, and reusable design assets.
Created future vision roadmaps to define where Epic’s AI should go next.
This gave me both a seat at the strategy table and ownership of the systems that made execution possible.
Disclaimer: Due to IP restrictions, actual Epic designs and assets cannot be shared, unless they've been made public.
Disclaimer: Due to IP restrictions, actual Epic designs and assets cannot be shared, unless they've been made public.
Epic, the largest EHR provider in the U.S., needed a way to make AI safe, consistent, and adoption-ready across 80+ clinical apps.
I built the company’s first AI design system from scratch, turning siloed AI projects into a unified, trusted foundation now used by 80K+ clinicians and 330+ healthcare organizations.
The Results
The AI design system became Epic’s foundation for all AI-driven experiences:
Trusted by 80K+ clinicians across 330+ healthcare orgs.
Rolled out before Epic’s largest annual conference, with 40,000+ attendees seeing the system in action.
Endorsed by Epic’s C-suite as a “beautiful” solution that aligned speed with adoption.
Established validated design standards that continue to guide all future AI initiatives.
"Working with Josh was a fantastic experience. He was highly responsive to feedback and played a key role in shaping our ideas into a compelling narrative and an enhanced user experience around AI.
I especially appreciated his initiative in researching the best solutions and his creativity in exploring new directions."
Nick Marzotto
Head of AI Customer Success, Epic
I used a design thinking approach with rapid prototyping, iterative feedback, and constant user validation to ensure we were building the right design system for real-world traction.
Step 1: Define the foundation
Audited 30+ in-progress AI features to uncover inconsistencies.
Identified key adoption blockers: lack of trust, unclear ownership, and no shared UX standards.
Step 2: Create reusable AI patterns
Defined 4 core AI workflows (summarization, task automation, content transformation, drafted text).
Built reusable components for explainability, trust signals, and safety checks.
Step 3: Establish visual + UX identity
Designed AI-specific icon system, visual language, and the “Bloom” symbol as the universal AI marker.
Standardized interaction rules across all apps.
Step 4: Scale + validate with users
Rolled system out across 80+ apps with dev + design teams.
Tested continuously with clinicians, iterated for safety and clarity.
I used a design thinking approach with rapid prototyping, iterative feedback, and constant user validation to ensure we were building the right design system for real-world traction.
The Process
Solving Hard Problems at Scale
While the actual design artifacts are confidential IP, these examples show the kinds of complex UX problems I solved, where design decisions directly impacted adoption, trust, and safety:
AI Design System
Before: Each team built AI their own way. Different layouts, controls, and voices. No shared framework. Good ideas got lost, bad ones repeated. Chaos at scale.
After: I created a unified AI design system spanning 80+ clinical apps. Teams reused components, workflows, and safeguards instead of reinventing. This gave Epic a shared language for AI and turned scattered experiments into a scalable system.
Trust & Explainability
Before: AI outputs showed up with no context. Clinicians didn’t know how reliable they were, so trust wavered.
After: I designed explainability patterns, such as confidence scores, source transparency, and clear controls (accept, adjust, dismiss). Clinicians could see why the AI suggested something and act with confidence.
Visual Identity
Before: AI had no “face.” It blended into Epic’s UI, so users couldn’t tell what was human vs. AI. Misinterpretation was a constant risk.
After: I partnered with designers to create a distinct AI identity called Bloom. Every AI output carried a consistent marker. Clinicians quickly learned to spot and interpret AI suggestions without confusion.
Scalability
Before: Every new AI feature was treated as a one-off pilot. Teams debated the same questions again and again. Progress slowed.
After: I shifted the mindset from “designing features” to “designing systems.” Built repeatable workflows, patterns, and safeguards that scaled. Teams moved faster, with trust and usability baked in from the start.
Broader Scope of Work
Beyond hands-on design, my role covered the larger ecosystem of AI at Epic:
This gave me both a seat at the strategy table and ownership of the systems that made execution possible.
Partnered with leadership to shape AI strategy across the company.
Built pitch decks and roadmaps used with partners (e.g. Microsoft) and major health systems.
Acted as an internal “AI design consultant” for 80+ teams and senior leadership.
Delivered presentations to R&D, Mayo Clinic, and executive audiences to educate and evangelize AI design.
Built and managed systems, like Figma component libraries, internal guidelines, and reusable design assets.
Created future vision roadmaps to define where Epic’s AI should go next.
Outcomes and Impact
This work established a unified AI design system that reshaped how 80+ clinical applications at Epic delivered AI. Instead of scattered features, teams could now scale safely through shared components, visual identity, and built-in safeguards. The result was faster development, reduced duplication, and consistent adoption across critical healthcare workflows.
Beyond usability, the design system influenced Epic’s broader AI strategy. It anchored partner presentations with Microsoft and Mayo Clinic, positioned design as the “AI expert” voice for leadership, and laid the foundation for future vision roadmaps. The impact extended from daily clinician workflows to enterprise-level partnerships, proving that UX can drive both adoption and strategic direction at scale.
Lessons for Founders
The patterns from this project extend directly to AI startups building their first or next product:
Think in systems, not features: Designing reusable patterns and safeguards pays off, whether you’re scaling from 1 product to 10 or 10 to 100.
Build trust into the workflow: Adoption hinges on explainability and user control. Treat them as core features, not extras.
Position for impact: A strong design system can influence partnerships, customer buy-in, and the credibility of your AI strategy.
This work proved how UX can turn fragmented AI pilots into a scalable system that clinicians trust.
At Epic, these design decisions shaped strategy, partnerships, and the path for how AI enters healthcare safely.
Outcomes and Impact
This work established a unified AI design system that reshaped how 80+ clinical applications at Epic delivered AI. Instead of scattered features, teams could now scale safely through shared components, visual identity, and built-in safeguards. The result was faster development, reduced duplication, and consistent adoption across critical healthcare workflows.
Beyond usability, the design system influenced Epic’s broader AI strategy. It anchored partner presentations with Microsoft and Mayo Clinic, positioned design as the “AI expert” voice for leadership, and laid the foundation for future vision roadmaps. The impact extended from daily clinician workflows to enterprise-level partnerships, proving that UX can drive both adoption and strategic direction at scale.
Lessons for Founders
The patterns from this project extend directly to AI startups building their first or next product:
Think in systems, not features: Designing reusable patterns and safeguards pays off, whether you’re scaling from 1 product to 10 or 10 to 100.
Build trust into the workflow: Adoption hinges on explainability and user control. Treat them as core features, not extras.
Position for impact: A strong design system can influence partnerships, customer buy-in, and the credibility of your AI strategy.
This work proved how UX can turn fragmented AI pilots into a scalable system that clinicians trust.
At Epic, these design decisions shaped strategy, partnerships, and the path for how AI enters healthcare safely.