JPMC Supplier Management Dashboard
My Role: Sr. Associate UX Designer
Duration: 4 Months
Tools Used: Figma Make, LLM
Project Overview
Due to the sensitive nature of this cybersecurity project, only limited details are shared here.
The Supplier Management Dashboard was designed for cybersecurity and third-party risk teams to centralize and streamline how supplier risk is investigated, monitored, and acted upon across a complex enterprise environment. The goal was to replace fragmented, manual workflows with a search-first experience that enables faster access to supplier intelligence, clearer risk visibility, and more efficient decision-making.
I led UX discovery and design, conducting user interviews and facilitating a cross-functional workshop to understand user workflows, pain points, and business needs. These insights informed a set of UX strategies focused on simplifying complex enterprise processes across information discovery, incident response, and reporting.
The experience was prototyped in Figma Make, with AI-assisted workflows used to rapidly translate research insights into structured interface concepts. Throughout the process, I worked closely with engineering to ensure feasibility within the constraints of the QlikSense platform, aligning design decisions with backend data structures and system integrations.
Explore in Figma
Research Summary
The research phase included 7 user interviews with key analyst personas and 1 facilitated workshop with cross-functional stakeholders to understand how supplier risk management was currently being executed across the organization.
My findings showed that users were operating in a highly manual environment, relying on Excel-based tracking and fragmented systems to manage supplier data, incidents, and engagements. This resulted in inconsistent information, duplicated effort, and slow access to critical risk insights.
Across all personas, the primary challenges were consistent: difficulty discovering and connecting supplier information across tools, limited visibility into overall risk posture, and inefficient workflows for investigation and reporting.
These insights directly informed the design direction for a search-first Supplier Management Dashboard, focused on centralizing supplier intelligence, reducing manual effort, and enabling faster, more informed decision-making.
Through stakeholder interviews and user research, I uncovered these pain points:
Manual Spreadsheet Tracking
Teams relied on manually updated Excel spreadsheets, creating inefficiencies, version control issues, and inconsistent supplier data visibility.
Fragmented Investigation Workflows
Critical supplier risk information was spread across multiple tools and sources, slowing investigations and decision-making.
Inconsistent Reporting Processes
Generating reports for leadership, governance reviews, and audits was time-consuming and lacked standardized workflows.
User Pain Points
How AI was Leveraged
This project used the firm’s LLM and Figma Make as a tightly integrated workflow to move from research insights to structured, high-fidelity UX prototypes at speed. Rather than using AI for one-off outputs, I treated LLM as a prompt engineering and systems design partner to translate messy research findings into precise, structured UI requirements that could be reliably executed in Figma Make.
User research insights and workshop outputs were first broken down into clear components such as user goals, primary tasks, interface hierarchy, and interaction rules. These were then converted into carefully curated, “design-ready” prompts in LLM, iteratively refined to enforce structure (e.g., layout hierarchy, component behavior, navigation patterns, and edge cases) rather than vague descriptions. Multiple prompt variations were tested to ensure consistency, clarity, and alignment with enterprise UX patterns before being passed into Figma Make.
Once refined, these prompts were used in Figma Make to generate initial wireframes and interface concepts, which were then iterated further by adjusting prompt specificity (e.g., refining component behavior, layout constraints, and data relationships) rather than manually redrawing from scratch. This allowed rapid exploration of multiple UX directions while maintaining structural consistency across screens. The result was a highly iterative, prompt-driven design process where AI functioned as a structured translation layer between research insight and executable interface design.