Discovery

Discovery

The architecture

The architecture

Design decisions

Design decisions

Reflection

Reflection

Designing a governed wealth management system — portfolio rebalancing as a decision architecture

Informed by 7 years of building governed wealth management systems at Y TREE — where I led the design of portfolio rebalancing tools managing over £9b in client assets.

This case study presents a portfolio rebalancing system for advisors managing UHNW clients across UK, Italy, and France, where every decision must be explainable, auditable, and mandate-aligned. Specific client data and proprietary product details have been abstracted.

Mar 2026 | Research, UX and UI design

Discovery

Problem framing

This isn't a dashboard problem. It's a decision governance problem.

The company operates across three markets with a mix of advisory and discretionary mandates. When portfolios drift, advisors don't just need to see the data — they need to act on it, justify it, and prove it was the right call.

The real design challenge was building a system that supports structured decision-making under regulatory pressure, not one that simply visualises numbers.

Problem framing

This isn't a dashboard problem. It's a decision governance problem.

The company operates across three markets with a mix of advisory and discretionary mandates. When portfolios drift, advisors don't just need to see the data — they need to act on it, justify it, and prove it was the right call.

The real design challenge was building a system that supports structured decision-making under regulatory pressure, not one that simply visualises numbers.

🧑‍🤝‍🧑

Multiple stakeholders

Advisors, UHNWIs, compliance officers, and portfolio managers with conflicting priorities

🧭

Complex workflow

First impressions are neutral, but the interface fails to sustain engagement.

⚖️

Regulatory accountability

Every recommendation must be explainable and traceable under MiFID II

📂

Fragmented data

Inconsistent sources with uneven reliability across markets

Stakeholder landscape

Three roles. Three definitions of what "right" looks like.

Before designing any screen, I mapped the decision ecosystem. The key insight was that decision power, capital ownership, and execution risk sit in entirely different hands — which means the system needs to serve fundamentally different mental models simultaneously.

Stakeholder landscape

Three roles. Three definitions of what "right" looks like.

Before designing any screen, I mapped the decision ecosystem. The key insight was that decision power, capital ownership, and execution risk sit in entirely different hands — which means the system needs to serve fundamentally different mental models simultaneously.

Advisors — The Decision Architect

shallow focus photo of woman face

“When allocations shift, I need clarity — not just alerts. Every adjustment must reinforce the long-term strategy.”

Pain: Too much time in Excel. No shared tool for tax impact. Hard to justify selling winners.

UHNWIs — The Informed Principal

shallow focus photo of woman face

“My wealth is a legacy. I need to understand how decisions today protect tomorrow."

Pain: 50-page PDFs that don't answer "Am I still safe?” Fear that rebalancing triggers unexpected tax.

Compliance & Execution — The System

shallow focus photo of woman face

Every action must remain mandate-aligned, auditable, and traceable. The system is both a safeguard and a bottleneck.

Strategic scoping

Why rebalancing — and not something else?

The full wealth management system spans onboarding, reporting, tax structuring, and more. I used a prioritisation matrix to identify the highest-leverage entry point.

Portfolio rebalancing scored highest across all three dimensions: financial impact, cross-team dependency, and regulatory sensitivity. It was also the workflow where the current process was most fragmented — advisors were managing drift detection and suitability reporting manually in Excel.

Strategic scoping

Why rebalancing — and not something else?

The full wealth management system spans onboarding, reporting, tax structuring, and more. I used a prioritisation matrix to identify the highest-leverage entry point.

Portfolio rebalancing scored highest across all three dimensions: financial impact, cross-team dependency, and regulatory sensitivity. It was also the workflow where the current process was most fragmented — advisors were managing drift detection and suitability reporting manually in Excel.

The architecture

Workflow

From drift to decision — an 8-step governed workflow

The architecture maps the full rebalancing journey: from automated drift detection through diagnosis, path planning, compliance validation, advisor review, simulation, client approval, and trade fulfilment. Each step has a defined owner, a clear decision gate, and an audit trail.

Workflow

From drift to decision — an 8-step governed workflow

The architecture maps the full rebalancing journey: from automated drift detection through diagnosis, path planning, compliance validation, advisor review, simulation, client approval, and trade fulfilment. Each step has a defined owner, a clear decision gate, and an audit trail.

Drift detection must enable 5-second awareness

The risk rebalancing dashboard surfaces portfolio-level alerts with immediate context — mandate breach vs. out-of-range vs. within band, sorted by impact severity. Advisors can see £1.5b AUM at risk across 298 clients at a glance, and act on the highest-priority cases without any manual scanning.

Design decision

Separating "Mandate breach" from "Out of range" wasn't just a labelling choice — it defines two different response protocols. Mandate breach requires immediate action; out of range requires monitoring. The visual hierarchy enforces this distinction.

Drift detection must enable 5-second awareness

The risk rebalancing dashboard surfaces portfolio-level alerts with immediate context — mandate breach vs. out-of-range vs. within band, sorted by impact severity. Advisors can see £1.5b AUM at risk across 298 clients at a glance, and act on the highest-priority cases without any manual scanning.

Design decision

Separating "Mandate breach" from "Out of range" wasn't just a labelling choice — it defines two different response protocols. Mandate breach requires immediate action; out of range requires monitoring. The visual hierarchy enforces this distinction.

Risk Intelligence

Once a drift is flagged, the advisor needs more than a number. This screen breaks down the risk score into its structural drivers — equity beta, growth concentration, USD exposure — and surfaces the top five contributing assets with their specific drift and impact. The AI rebalance suggestion appears inline, alongside the strategic trigger and expected impact, so the advisor can move from diagnosis to action without switching context.

Design decision

The risk breakdown and the AI suggestion are deliberately placed on the same screen. In the previous workflow, advisors had to cross-reference multiple tools to understand why a portfolio drifted and what to do about it. Collapsing diagnosis and recommendation into a single view reduces the time between detection and action — and makes the advisor's rationale easier to construct when they need to justify the decision to compliance or the client.

Risk Intelligence

Once a drift is flagged, the advisor needs more than a number. This screen breaks down the risk score into its structural drivers — equity beta, growth concentration, USD exposure — and surfaces the top five contributing assets with their specific drift and impact. The AI rebalance suggestion appears inline, alongside the strategic trigger and expected impact, so the advisor can move from diagnosis to action without switching context.

Design decision

The risk breakdown and the AI suggestion are deliberately placed on the same screen. In the previous workflow, advisors had to cross-reference multiple tools to understand why a portfolio drifted and what to do about it. Collapsing diagnosis and recommendation into a single view reduces the time between detection and action — and makes the advisor's rationale easier to construct when they need to justify the decision to compliance or the client.

AI recommends. Humans decide.

The advice tool generates AI-recommended buy/sell adjustments based on the All-Weather model. But advisors retain full discretion — they can amend individual positions, add a counter view, or regenerate the proposal entirely.

Design decision

The AI suggestion is surfaced as a starting point, not a default. The "Counter view" and "Rationale" buttons are primary-level actions — not buried — because overriding the AI should feel supported, not like breaking the system.

AI recommends. Humans decide.

The advice tool generates AI-recommended buy/sell adjustments based on the All-Weather model. But advisors retain full discretion — they can amend individual positions, add a counter view, or regenerate the proposal entirely.

Design decision

The AI suggestion is surfaced as a starting point, not a default. The "Counter view" and "Rationale" buttons are primary-level actions — not buried — because overriding the AI should feel supported, not like breaking the system.

Governance by design — compliance embedded, not bolted on

Compliance checks run automatically as the advisor builds the proposal. Rather than a separate approval gate at the end, the system validates mandate alignment, restricted securities, suitability rules, concentration limits, and liquidity requirements inline.

Design decision

Making compliance visible during construction — not just at submission — changes advisor behaviour. It turns compliance from a blocker into a guardrail.

Governance by design — compliance embedded, not bolted on

Compliance checks run automatically as the advisor builds the proposal. Rather than a separate approval gate at the end, the system validates mandate alignment, restricted securities, suitability rules, concentration limits, and liquidity requirements inline.

Design decision

Making compliance visible during construction — not just at submission — changes advisor behaviour. It turns compliance from a blocker into a guardrail.

From advice to client suitability — one flow, not two

Once the advisor finalises the proposal, the system generates a client-facing suitability report with AI-drafted rationale. The advisor reviews and edits before sending. The client receives a mobile-optimised view with proposed actions, stress test scenarios, and a clear Approve / Discuss / Decline decision.

Design decision

The suitability report and the internal advice tool share the same data model. This eliminates the manual re-keying that was the biggest source of error and delay in the legacy process.

From advice to client suitability — one flow, not two

Once the advisor finalises the proposal, the system generates a client-facing suitability report with AI-drafted rationale. The advisor reviews and edits before sending. The client receives a mobile-optimised view with proposed actions, stress test scenarios, and a clear Approve / Discuss / Decline decision.

Design decision

The suitability report and the internal advice tool share the same data model. This eliminates the manual re-keying that was the biggest source of error and delay in the legacy process.

From approved advice to executed trades — closing the governance loop

Once the client approves the rebalancing proposal, the system generates the corresponding trade orders and routes them into the trading operations queue. This screen gives the back-office team a real-time view of all in-progress trades across clients — with status tabs for pending approval, awaiting scheduling, scheduled, due for execution, and order submitted. Each trade row carries the full context needed for execution: instrument, ISIN, action, amount, tax wrapper, approval date, and requested execution window.

Design decision

The status pipeline at the top isn't just a summary — it's the primary navigation affordance. Back-office operators work through trades sequentially by status, not by client. Surfacing the count at each stage lets them triage at a glance without opening individual records. This mirrors how trading desks actually operate: by queue priority, not by relationship.

From approved advice to executed trades — closing the governance loop

Once the client approves the rebalancing proposal, the system generates the corresponding trade orders and routes them into the trading operations queue. This screen gives the back-office team a real-time view of all in-progress trades across clients — with status tabs for pending approval, awaiting scheduling, scheduled, due for execution, and order submitted. Each trade row carries the full context needed for execution: instrument, ISIN, action, amount, tax wrapper, approval date, and requested execution window.

Design decision

The status pipeline at the top isn't just a summary — it's the primary navigation affordance. Back-office operators work through trades sequentially by status, not by client. Surfacing the count at each stage lets them triage at a glance without opening individual records. This mirrors how trading desks actually operate: by queue priority, not by relationship.

Key design decisions

Three decisions that shaped the system

Three decisions that shaped the system

🧑‍⚖️

Decision 1

Why not full automation?

The system could technically execute rebalancing without advisor input. We chose not to. In a regulated advisory context, automation without oversight creates liability — and erodes the trust relationship between advisor and client. The design keeps humans in the loop at every consequential step.

🚦

Decision 2

Why embed compliance inline, not as a final gate?

A compliance check at the end of a workflow is a blocker. A compliance check during construction is a tool. Moving validation inline reduced proposal rejection rates and gave advisors confidence to iterate faster.

🔗

Decision 3

Why treat the suitability report as part of the same tool?

Previously, advisors would build a recommendation in one system and re-create the client communication in another. Unifying these into a single flow eliminated a critical failure point and reduced preparation time significantly.

Reflection

What I learned

The hardest part wasn't the UI — it was defining what "done" looks like

In a system this complex, the design work isn't finished when the screens look right. It's finished when every stakeholder knows exactly what they're responsible for, and the system supports that clarity without ambiguity.

The biggest challenge was designing for a workflow where an error doesn't just mean a bad experience — it means a regulatory breach, a financial loss, or a breakdown in client trust. That constraint made every decision more deliberate, and ultimately produced a more considered system.

What I learned

The hardest part wasn't the UI — it was defining what "done" looks like

In a system this complex, the design work isn't finished when the screens look right. It's finished when every stakeholder knows exactly what they're responsible for, and the system supports that clarity without ambiguity.

The biggest challenge was designing for a workflow where an error doesn't just mean a bad experience — it means a regulatory breach, a financial loss, or a breakdown in client trust. That constraint made every decision more deliberate, and ultimately produced a more considered system.

Chun-Chuan Lin

Say hello at designlcc@gmail.com

Built by ux-circles.com

Chun-Chuan Lin

Say hello at designlcc@gmail.com

Built by ux-circles.com

Chun-Chuan Lin

Say hello at designlcc@gmail.com

Built by UX CIRCLES

Create a free website with Framer, the website builder loved by startups, designers and agencies.