ERP - AI in Design
A story about how I Survived Six 0→1 ERP Projects in One Month
Overview
Epic Systems is a leading healthcare software provider, specializing in electronic health records (EHR) and integrated health information systems. Its solutions help hospitals and clinics manage patient care, clinical operations, and administrative tasks seamlessly.
I leveraged AI and vibe coding to integrate various AI tools into design workflows, accelerating productivity while balancing input from human experts & user experience designers. This approach ensured designs remained empathetic, strategic, and human-centered, carefully considering end-user needs.
Timeline
2025
Role
Leader Product Designer
Skills
Information Architecture
User Flows
Feature Scope
Note
This article shares my personal experience as the sole product designer working on complex B2B systems in the era of AI. Due to confidentiality, I won’t delve into specific features or product details.
PROBLEM
Surface AI-driven highlights for actionable next steps
“Hey Monica, we need to do this, and this, and also this…”
When I first joined the billing financial experience team, I never imagined that within just half a year of experience as a junior product designer, I would own an entire system. But that's how a fast-paced company operates.
One day, our product lead called a meeting and casually dropped, “Hey Monica, we need to do this, and this, and this. We discussed it with customers, and they're excited, so we really need something ready by next month.....”
Problem..?
No prior domain financial knowledge, no existing product, no UX precedent. As the sole designer working with a team of developers, I faced building the most complex type of B2B software imaginable: an ERP system for healthcare finance.
Design in 2025 and forward
That is, until I stopped designing like it was 2020 and began thinking as if it were 2030—not by trying to handle everything alone, but by partnering strategically with AI and leveraging it to do what it does best, while I focused on what AI couldn’t.
Before starting points
The Reality of Designing for Complexity
“Hey GPTs, I don't know this!"
After conducing user interviews and talk with end users during immersion, I found that it shouldn't be the user's job to surface clarity and confidence fast for data table.
In our case, none of these elements were documented. We were building everything from scratch, requiring me to understand each component deeply enough to design workflows that actually made sense. Traditional UX methods fell short; there simply wasn't time for lengthy research cycles, and user access was limited. I had to accelerate my learning curve quickly.
Using chatgpt to understand background
Part I: design with ai
Use Ai to amplify
AI is a highly capable, direction-hungry design intern.
Early on, I shifted my mindset. Rather than treating AI merely as a shiny new tool for auto-layouts or generating icons. AI couldn't grasp the nuances of our product vision or business logic, but it significantly accelerated information processing and idea generation, allowing me more time to focus on strategic thinking.
break design process with ai
Part II: designer-ai
Building a Designer-AI Workflow
Training AI with Context
I treated AI as a learner, feeding it comprehensive meeting notes, user research data, and even whiteboard photos. In return, it summarized information and highlighted key themes rapidly.
Developing Design Artifacts with AI
Instead of starting from blank Figma frames, I used AI to quickly draft rough structures based on user personas, journey maps, and pain points. This provided a solid foundation on which to apply human judgment and refinement.
Rapid Iteration and Prototyping using AI-Driven Prototyping
I collaborated closely with engineers and stakeholders, leveraging Figma Make to rapidly prototype interactive demos—completely skipping traditional wireframes. Within a single week, we iteratively co-designed over six workflows.
They weren't pixel-perfect, but they provided clear direction and functionality.
feature prototype
part iii: Research & iteration
Where Most AI Tools Fall Short (And How I Avoided It)
Great design doesn’t start with prompts.
There’s a common misconception that AI can entirely replace designers. Many tools promise “prompt-to-app” solutions—simply type "Design an ERP System" and get a polished product. However, this is misleading and overly ambitious.
Good outputs starts with clarity: business goals, user needs, edge cases, technical constraints.However, this is misleading and overly ambitious.
If your input is shallow, your output inevitably will be too. AI requires careful guidance, thorough context, and constant feedback to deliver valuable results. Treating AI as a junior teammate allowed me to delegate specific tasks effectively while maintaining overall design quality.
prompts & design
outcomes
Six Wins in Two Weeks, One Designer
By the end of two intense weeks, I successfully produced six fully tested interactive workflow demos, each covering critical ERP functionalities. Every demo underwent rigorous testing and thorough documentation, all achieved without burnout thanks to AI handling heavy lifting tasks, enabling me to focus on strategic decision-making.
Results
Most importantly, my perspective shifted from asking “How can AI design for me?” to exploring “How can I design better with AI?”
reflection
live with ai tools...
AI Won’t Replace Designers - But Designers Who Use AI Will Replace Those Who Don’t?
If you’re still designing in a Figma Frame-only world, you’re missing the opportunity. AI won’t replace you — but it will reshape the skill set that makes you valuable.
Your job isn't to push pixels; it's to guide intelligence—both human and machine—toward solving the right problems for the right users in the right ways.
Stay human
In reality, AI is your smartest teammate—waiting for clear, strategic direction, but AI can mimic tone — but not empathy, judgment, or vision.
Designer & developer
When developers see AI, they recognize speed. When designers see AI, many perceive risk. Just work with AI!
