Back to Projects
DialogueIQ – Turning Client Conversations into Growth
Wealth Management
Multi-Agent Systems
AI Assistant
Client Advisory
OpenAI
Supabase
TypeScript
Hackathon
Zurich AI Hackathon – UBS Challenge (Zurich, Switzerland) • September 2025
Product & AI Lead
Project Overview
Built for the UBS challenge at the Zurich AI Hackathon, DialogueIQ helps wealth advisors reclaim time from admin and reinvest it into high-value client conversations. It ingests meeting transcripts, updates systems automatically, and uses a multi-agent setup to surface a prioritized task list with clear rationales. Advisors see exactly which clients to follow up with, why, and how—turning everyday dialogue into measurable growth.
Challenges
- •Translating messy, unstructured client conversations into structured insights and actions advisors can trust
- •Designing a dashboard that highlights the most important client opportunities without overwhelming the advisor
- •Coordinating multiple OpenAI-powered agents for extraction, classification, and prioritization while keeping latency low
- •Modeling realistic wealth-management data and workflows without direct access to UBS production systems
- •Ensuring that AI-generated recommendations remain interpretable and aligned with regulatory and compliance constraints
Key Achievements
- •Delivered an end-to-end prototype that turns conversation transcripts into a prioritized, advisor-ready task list
- •Implemented a multi-agent orchestration flow using the OpenAI API for insight extraction and prioritization
- •Built a v0/TypeScript frontend that visualizes client profiles, insights, and AI-recommended actions in a single view
- •Persisted conversations, client data, and recommended actions in Supabase to support iterative advisor workflows
- •Deployed the full stack on Vercel, enabling judges to explore the live dashboard experience during the hackathon demo
Technologies Used
TypeScript
v0 (Frontend)
OpenAI Multi-Agent Orchestration
Supabase
Vercel