Dramatically
Faster Insights
Time to get risk report
Loading...
Lake Bluff, IL/
Commercial Risk Management
Gen AI Risk Register Assistant
Role
Design Lead
Timeline
2024 - Present
Team
Data Science, Risk Officers, Frontend Engineers
Tools
Figma, LLM Design Patterns, Audit Compliance, AI/LLM Platform, React, Python Backend
Executive Summary
The Risk Register Assistant is a Gen AI-powered chatbot designed to democratize access to complex risk data. It enables risk officers to instantly query executive summaries, regulatory readiness status, and peer benchmarks using natural language, all while maintaining a strict audit trail.
My Role
- •Designed the conversational UI specifically for risk querying workflows
- •Implemented 'Recommended Actions' to guide users to high-value insights
- •Designed the audit trail interface for regulatory compliance
- •Created data visualization cards for risk posture trends
- •Established governance patterns for AI responses
The Challenge
The Problem
Risk data was buried in complex spreadsheets and disparate dashboards. Executives and Risk Officers needed a way to get instant answers to questions like 'What is our current regulatory readiness?' without manual data crunching.
User Impact
Empowered risk officers to get instant answers to complex queries without manual data crunching.
Business Impact
Reduced time-to-insight for executive risk reporting from days to seconds, enabling proactive risk management.
Process
01
Governance
Established AI guardrails and audit requirements.
Key Activities
- •Risk Assessment
- •Governance Workshops
02
Design
Designed the conversational patterns and audit UI.
Key Activities
- •Prompt Engineering
- •UI Design
Solution
An intelligent assistant that sits on top of centralized risk data. The interface features a persistent history sidebar for context, a clean query interface, and 'Recommended Action' chips that surface rising risks. Every interaction is time-stamped and traceable for audit purposes.
Key Features
- •Conversational Interface: natural language queries for complex risk data
- •Audit Trail: full traceability of all inquiries and AI responses
- •Recommended Actions: proactive surfacing of rising risks and comparisons
- •Data Governance: role-based access control within the chat context
- •Rich Cards: integrated data visualizations within the chat stream
Results
100%
Traceability
Audit trail coverage
Design Artifacts
Chat interface with recommended actions and history
design
Reflection
What Worked Well
- Recommended actions guided users to value
- Persistent history provided crucial context
Challenges Overcome
- Hallucination risks in high-stakes environment
- Visualizing complex risk data within a chat stream
Key Takeaways
- Trust is the currency of Enterprise AI
- Conversational UI democratizes access to complex data