AI-Powered Investing

From Fragmented Advisory to AI-Driven Investment Decisions

Project Date: 2025 - 2026 , Zürich | Clients: UBS | My Role: Service Design

Problem framing

Advisors rely on 10+ internal platforms to access fragmented data and insights.

Core friction

Fragmented tools → constant context switching
Information overload → hard to synthesize quickly
Manual workflows → proposals rebuilt from scratch

Quantified impact

15–30 minutes lost per client interaction

Key Insights

Advisors don’t actively search — they scan and react
Value lies in decision acceleration, not information access
Fragmentation — not individual tools is the core bottleneck

Understanding the advisor workflow

Mapping the advisory workflow revealed that fragmentation impacts every stage — from discovery to client communication.

Design Principles

1. Design for scanning, not searching

Surface opportunities proactively instead of relying on manual queries

2. Optimize for decision-making, not information access

Prioritize actionable insights over raw information

3. Solve fragmentation through centralization

Unify tools into a single decision layer

Solution

Based on these principles, I contributed to shaping the Investment Hub as a centralized decision layer combining:

1. Signal → Opportunity surfacing

AI scans client portfolios and market signals to surface relevant investment opportunities.

2. Context → Intelligent matching

Links research insights to specific client portfolios and recommends suitable products.

3. Output → Client-ready generation

Generates clear, personalized investment messages ready for client communication.

End-to-end advisory workflow powered by AI

The redesigned workflow integrates AI capabilities into each stage — enabling end-to-end decision-making without context switching.

Each stage is powered by integrated modules such as Next Best Client, Investment Hub, and Storyteller, enabling end-to-end decision-making without context switching.

Product Architecture

To support this workflow, I defined a modular architecture that connects AI capabilities with advisory workflows.

The platform architecture illustrates how global AI capabilities, structured workspaces, and embedded insight modules come together to support the end-to-end advisory workflow.

It is structured around three key layers:

Global AI capabilities (e.g. search, summarization)
Structured workspaces aligned with advisor tasks
Embedded insight modules supporting decision-making

Together, these form a unified decision layer that enables end-to-end advisory without context switching.

Explore supporting modules

1. Next Best Client

(Supports: Opportunity identification)

Identifies high-priority clients based on portfolio signals, enabling proactive opportunity discovery.

2. Storyteller (Portfolio Summary)

(Supports: Insight compression, Client-ready output)

Transforms fragmented data into concise, client-ready narratives to support communication.

3. Investment Hub

(Supports: Intelligent matching, Decision-making)

Serves as the central workspace where insights are aggregated and decisions are made.

4. OneSearch

(Supports: Information access across all stages)

Provides a unified search layer to access, compare, and evaluate information across systems. (Detailed case study in the following section.)

Example flows

Flow 1 — From opportunity discovery to client-ready proposal

Identify opportunity → Evaluate options → Generate proposal → Present to client

Flow 2 — From passive monitoring to proactive client prioritization

Scan signals → Identify high-priority clients → Take action proactively

Impact

The Investment Hub significantly reduces time spent searching and synthesizing insights:

5–10 minutes saved per client when gathering research and bank views
Up to 30 minutes saved for complex proposals involving product comparison

This allows advisors to focus more on client interaction and value delivery, rather than operational tasks.