Xylo is an AI-powered decision intelligence software that helps financial advisors identify relationship risks and opportunities in real time.
CONTRIBUTION
Lead Product Design
UX Research
TEAM
CEO
Design Manager
Engineers*3
TOOLS
Figma
Zeplin
Jira
TIMELINE
4 Months
Context
Redesign the Xylo Dashboard to Transform Client Monitoring into Actionable Intelligence
In the summer of 2025, I worked at Xylo AI as the Product Design Intern, leading the redesign of the core dashboard experience, aiming to transform it from a traditional CRM-style interface into a proactive decision system.
The primary challenge was not the lack of data, but the overwhelming volume of fragmented signals. Advisors were managing dozens of client relationships across multiple systems, yet struggled to quickly determine which relationships required attention.
To address this, I reframed the product from a “data dashboard” into a “relationship monitoring and prioritization engine.” Through behavioral research, workflow observation, and iterative prototyping, I introduced a triage-based system that surfaces changes, risks, and urgency signals in real time.
This strategic shift significantly reduced cognitive load and repositioned the platform from passive data storage to active decision support.
My Role
I Led the strategic redesign of the Dashboard architecture, redefining how relationship intelligence is structured, prioritized, and surfaced to advisors.
I Advanced core decision flows by shifting the information hierarchy from static client profiles to change-based monitoring and action prioritization.
Impact
Increased advisor workflow efficiency by 20% by reducing decision hesitation and manual signal analysis.
Improved prioritization clarity, leading to higher engagement with Top Actions and proactive outreach.
Understand
From Users: Information is abundant
but fragmented.
I conducted the user interviews with advisors, they described their workflow as mentally exhausting. They searched through emails, CRM notes, and call records to prepare for client conversations.They constantly switch between tools to answer three core questions: Who needs my attention right now? What has changed? What should I do next?
“The numbers are easy. The hard part is understanding what my clients need from me at the right moment and communicating it clearly.”
— Bob, Financial Advisor
Project Goal & Scope
Translate fragmented client signals into clear, confident actions, then reduce cognitive load.
Project Task 01
Redesigning the data dashboard experience into a relationship monitoring engine
Problems
The Dashboard Functioned as a Data Summary — Not a Decision Tool
The previous version of the Xylo Dashboard was designed as a summarized view of the existing CRM system. It consolidated metrics such as AUM, health scores, sentiment, fees, referral warmth, and contact history into a comprehensive table. While information was centralized, it did not translate into clarity.
The table contained too many static data points, requiring advisors to manually scan and interpret signals across multiple columns. Important changes, such as sentiment drops or engagement gaps were visually subtle and easy to miss.
Approach
Interviews and usability testing to identify true decision drivers and eliminate non-essential signals
To redesign the dashboard effectively, I first focused on understanding what actually drives advisors’ decisions rather than assuming which data points were important
I also conducted usability testing to validate the new prioritization logic
Design Decision
Prioritization by Change, Not by Status
I redesigned the dashboard to prioritize change-based signals rather than static summaries. Sentiment became “Sentiment ↓18% (7d)” instead of “87 (-18%).” Inactivity became a time-triggered alert. Risk states became visible clusters instead of subtle indicators.
However, surfacing change alone was not enough and signals needed to translate into action.
The original “Top Actions” module grouped alerts by category (Churn Risk, Opportunity, etc.), requiring advisors to interpret urgency themselves. I redesigned it around decision context instead of system taxonomy.
Each action now clearly answers: Who needs attention, Why now and How to act.
By adding explicit triggers (“47 days silence”) and suggested channels, the interface shifts from displaying alerts to guiding next steps by reducing ambiguity and accelerating decision-making.
Project Task 02
Evaluating the Personalized Relationship Intelligence
Our Initial Approach
Separating “Who the Client Is” from “How to Work With Them”
We hypothesized that placing too much information on a single page could increase cognitive load, making it harder for advisors to quickly locate the insights they need.
By separating different types of client insights, we aimed to make information easier to scan, understand, and act on.
For our first exploration, we divided insights into two pages: Personality Page and Relationship Page.
What We Found Out
Advisors don’t think in “personality vs. relationship.” They think in “How do I approach this client, right now?”
To validate our initial design assumptions, we conducted tree testing and usability testing with financial advisors.
These methods allowed us to evaluate whether our information structure aligned with advisors’ mental models during real decision-making tasks.
Design Decision
Combine the Relationship Page and Personality Page into a unified Client Dossiers experience
Because advisors think in terms of immediate approach and next action, we unified personality and relationship insights into a single view.
The Client Dossiers page focuses on browsing, filtering, and selecting clients, and the Client Profile page supports deep understanding, relationship context, and action planning.
Project Task 03
Proactive Signal Detection & Action Hub
Trust & Control in AI-Generated Messages
When using the AI draft generator in Xylo’s Top Actions, financial advisors lacked confidence in adopting AI-generated messages due to limited transparency and control.
Indicate why the draft was suggested
Real-time adjust writing tone
Constraints drive focus
Limited time and budget encouraged high-impact prioritization over complexity, reinforcing that an MVP is a starting point that evolves through continuous iteration.
B2B UX prioritizes efficiency
This project deepened my understanding of role-based workflows, decision-support design, and data visualization and core considerations in B2B SaaS products.
Organization enables collaboration
