AI-Powered (FinSight) Anomaly Detection in FinLink
Designing a smart, transparent alert system to flag unusual financial transactions and boost decision-making confidence.
Self-initiated
PLATFORM
Web Application for Desktop (FinLink)
Role
UX Design
UI Design
Research
Context
FinLink was designed to help users manage complex financial data with clarity. But during user research, one insight stood out: users often miss unexpected spikes, drops or irregularities in their transactions until it’s too late.
Traditional dashboards showed data, but didn’t warn users. So, we asked:
“What if FinSight (FinLink’s AI) could think like a
vigilant financial assistant?”
Goal
Design an AI-powered(FinSight) feature that detects transaction anomalies automatically and surfaces them proactively to users in a meaningful and contextual way.
Research Insights
From interviews with finance managers and business owners:
“I wish I’d caught that vendor double charge earlier.”
“Some of these spend patterns are hard to track manually.”
“There’s too much data, I don’t always know what to look for.”
This validated a clear opportunity to leverage AI for anomaly detection.
Approach
We broke the design challenge into two parts:
Detection
Identify transactions that deviate from past patterns using AI models.
Part 1
Delivery
Inform users when, where and how to take action without overwhelming them.
Part 2
We mapped the UX flow from
Backend Intelligence
UI Notifications
User Decision-making
UX Goals
Keep alerts actionable and non-intrusive
Provide clear explanations for anomalies
Let users validate, dismiss or flag anomalies
UX Flow
When FinLink’s AI (FinSight) spots an unusual transaction, a banner appears on the dashboard with options to Review, Dismiss, or Flag. Clicking Review takes the user to a detailed view explaining what happened and why. From there, they can mark it safe, flag it, or add a note, keeping users in control without breaking their workflow.
Design Decisions
The anomaly alert banner was placed above the dashboard cards to catch attention without being intrusive. Its colour signals urgency while blending with the overall UI. Primary, secondary and tertiary actions (Review, Flag and Dismiss) give users clear, immediate choices. We also added a subtle Explanation Toggle "Why are you seeing this?" link for transparency.
Outcome
This AI-powered anomaly alert became a standout feature in FinLink’s move towards smarter, data-driven experiences.
Reduced missed anomalies
Boosted user trust in the system
Saved time for financial teams previously combing through data manually
Conclusion
This project demonstrated how AI can enhance financial decision-making by flagging unusual transactions and offering users clear, actionable choices. Through thoughtful UX flows, transparent alerts and user-first interactions, FinLink not only made anomaly detection intuitive but also built trust by showing users why certain insights surfaced. This approach sets the foundation for smarter, AI-driven financial experiences ahead.