Turn your Design system into a design engine

TrueUI is a Figma plugin I conceived, designed, developed, and launched to help AI work reliably with existing design systems.

The product began with a simple observation: AI could generate visually convincing interfaces, but it repeatedly ignored or misused the system it was supposed to follow.

I initially approached this as a design-generation problem. It became clear that the more fundamental challenge was system understanding: what information AI needs, how that knowledge should be prepared, when the designer should intervene, and how generation should be validated.

TrueUI became a product that combines design-system mapping, human review, task-specific context preparation, native Figma execution, and validation.

Year

2026

Role

Founder, Product Designer, AI Workflow Architect
& Developer
Overview
The Problem
Step Zero
Mapping
AI Workflow
Generation
Components
Validation
Outcome

Copilot Main Features

LLM Chatbots
Admin chatbot enhances IT tasks, end-user chatbot streamlines self-service

Fine-tuning Tools
Refines chatbots for accuracy, reducing workload with independent query resolution.

Suggested Categories
AI suggests relevant ticket categories to streamline classification and routing.

Case Summarization
Generates concise ticket summaries with key details for quick review.

AI Emotion
Analyzes user sentiment to flag frustrated users and prioritize issues.

AI Author
Drafts context-specific responses to ensure clear, professional ticket communication.

Design Process

Challenges

One of the most significant challenges was designing pioneering features that were not only new to the ITSM industry but also groundbreaking on a global scale. 

  • Provide practical solutions for both end-users and admins without overwhelming them.
  • Balance automation with user control and clarity.
  • Be adaptable across different departments (e.g., IT, HR, Finance).
  • Set a standard for AI-driven efficiency in an emerging market.

My Role – Product Design Lead

As the Lead Product Designer, I managed a small team of two additional product designers, guiding the project through all phases of the design process. 

  • Leading the design and execution of AI-driven tools.
  • Overseeing prototyping, A/B testing, and design iterations based on customer feedback and market research.
  • Ensuring cohesive collaboration with product managers, developers, and stakeholders.
  • Driving the creation of an intuitive, user-centric interface while ensuring the features were innovative and met the needs of both admins and end-users.

Research & Discovery

The research phase shaped SysAid CoPilot’s AI tools through feedback and market analysis, refining key features to address user needs.

Design Partner Feedback

Collaboration during alpha & beta testing provided insights to improve usability and feature alignment
Pain Points: Discoverability issues and blockers like admin chatbot placement.
Refinements: Weekly testing refined features like AI ticket summarization and streamlined workflows.

Customer Feedback

Feedback from surveys, forms, and usage analytics revealed key improvement areas:
Pain Points: Limited chatbot customization and discoverability issues.
Positive Feedback: Led to Fine Tune tools for admins and optimized chatbot responses.

User Personas

Research identified three user groups with distinct needs and challenges:

  • End Users: Struggle with limited self-service options and slow resolutions.
  • Admins: Face manual categorization issues and limited AI customization.
  • System Admins: Need advanced customization but struggle with tracking AI’s long-term impact.

Task Complexity and AI Engagement by User Role

Conclusion: Design Solution

The design solutions focused on simplifying workflows while supporting advanced functionality:

  • Tailored AI Tools: Role-based features for task-specific efficiency.
  • Minimalistic Layout: Clear, clutter-free interface for quick actions.
  • Consistent Design: Familiar visual hierarchy for better navigation.
  • Flexible Functionality: Tools adaptable to user skill levels.
  • Seamless Integration: Enhances workflows without overwhelming users.

LLM Chatbots

Admin Chatbot

The admin chatbot in SysAid streamlines routine IT tasks like ticket resolution and solution generation, significantly reducing manual workload. Using natural language processing, admins can quickly scan tickets, apply solutions, and generate guides, enhancing efficiency.

Design Strategy

The chatbot was designed as an overlay panel to balance visibility with workflow continuity. Accessible AI recommendations and automation require minimal input, enabling admins to complete tasks rapidly while still supporting advanced, complex workflows.

Resolving a Service record

This video demonstrates how SysAid’s AI-powered chatbot streamlines service record resolution by identifying urgent tickets, suggesting solutions, and generating step-by-step guides, automating tasks to improve efficiency and response times.

End-User Chatbot

The AI Chatbot for End Users simplifies issue reporting and self-service, allowing employees to bypass IT teams for routine inquiries or service requests. Designed for ease of use, it connects directly to SysAid’s ITSM resources, offering a streamlined way for users to troubleshoot common issues or escalate unresolved problems quickly and independently.

Design Strategy

We designed a familiar interface using popular chatbot patterns to ensure a minimal learning curve. The AI integrates with SysAid’s Self-Service Portal and third-party tools, allowing users to resolve issues or submit service records within their preferred environments.

Settings & Tuning

Efficient big data handling allows system admins to refine chatbots for consistent, accurate answers, reducing IT workload by independently resolving user queries.

Design Strategy
Data Weighting: Prioritizes key data sources.
Multi-Chatbot Customization: Tailors bots for specific departments.
Minimal Overwhelm: A clean, intuitive UI prevents data overload.

Usage Dashboard

The AI Usage Dashboard provides a comprehensive overview of system utilization, allowing admins to track both end-user and admin interactions with the AI. Key metrics such as service records touched by AI, categorized interactions, and the impact of internal data are displayed, offering valuable insights that support effective decision-making and AI optimization.

Design Strategy

Data Clarity
Key metrics are prominently displayed for easy, quick analysis.

Actionable Insights
Visual cues highlight trends and deviations, aiding swift decision-making.

Customization Flexibility
Flexible filters and views allow admins to tailor the dashboard to their needs.

Data Pool

The Data Pool consolidates various knowledge sources, such as service records, documents, and knowledge bases, into one central location for each chatbot. This ensures the AI pulls relevant information tailored to specific queries, minimizing reliance on IT staff while maintaining chatbot accuracy.

Design Strategy

Color-Coded Chips
Unique colors and icons for each data source enable quick recognition.

Visual Weight System
A 1-5 ball system displays each source’s importance at a glance.

Split-Screen Layout
Selecting a source opens a split-screen for easy content management

Monitor & Fine-tuning

Allows admins to continuously improve the chatbot’s performance by reviewing the quality of its responses. Every question and corresponding answer is displayed in an organized table, providing transparency into how the chatbot functions. Admins can modify, approve, or flag answers for future refinement to ensure the accuracy and reliability of chatbot responses.

Design Strategy

Editable Responses
Admins can refine answers directly for improved accuracy.

Quality Scoring
Each response has a quality score to highlight improvement areas.

Visual Cues
Color-coded indicators provide quick insights into response quality.

Refined Interaction
Quick actions, like marking answers as irrelevant, streamline the process.