Auto Audit

Auto Audit

Auto Audit

A platform that collaborates with AI tools to support compliance and risk-review workflows at scale.

A platform that collaborates with AI tools to support compliance and risk-review workflows at scale.

A platform that collaborates with AI tools to support compliance and risk-review workflows at scale.

Please note: some details have been generalized to respect confidentiality agreements.

Please note: some details have been generalized to respect confidentiality agreements.

Please note: some details have been generalized to respect confidentiality agreements.

INDUSTRY

INDUSTRY

Banking

B2B

Internal Tools


ROLE

UX/UI Design

COLLABORATED with

Product Owner (1)

Researcher (1)

Engineering Team (12)

TOOLS

TOOLS

Figma

Sketch

Figma

Sketch

Figma

Sketch

DURATION

DURATION

8 months

(Mar. 2022 – Nov. 2022)

8 months

(Mar. 2022 – Nov. 2022)

8 months

(Mar. 2022 – Nov. 2022)

Background

Risk Analysts review customer actions like large transfers or profile changes. A new AI audit engine was planned to be introduced to run initial checks, but many cases still require human review and remediation.

Problem

AI runs first-pass checks, but unresolved cases still require analysts to investigate, judge, and act.

The challenge was not just scaling audits, but redesigning how AI and analysts collaborate to resolve risks across roles.

Business goal

"Enable 100% audit coverage using advanced analytics, and reduce review time, while freeing staff to focus on complex support needs."

Solution

A role-based audit platform that integrates AI first-pass checks with human-led investigation and remediation, designed to support clear decisions, ownership, and cross-role collaboration.

Context

System map

When AI flags an issue, people take over

The system runs an initial audit and auto-allocates unresolved cases to analysts. From there, multiple teams step in: checking, remediating, reallocating, until each case is resolved or passed back to the system.

Context

The Challenge

System Complexity Across Roles & Channels

While AI now handles the initial act, the human side of the system remained unchanged.
Cases that couldn’t be auto-resolved were passed back into existing manual workflows, where Risk Analysts, Operation Analysts, and Team Managers still worked across disconnected tools, channels, and inconsistent review criteria.

Roles

RA (Contact Centre)

RA (Branch)

Operation Analyst

Team Managers

Primary task

Review

Review

Remediate

Manage

Primary channel

Voice

Documents

Documents

Coordination

Tools Used

  • Call Review system

  • Excel sheet

  • Branch document

  • Note tracker

  • Internal manual operating system

  • Excel Dashboard

  • Email

The challenge wasn’t just fragmented tools, it was fragmented perspectives.

Research

User Problems

Every role engages with the system differently, and every disconnect impacts the outcome. Through conversations with 8 users across risk analysis, remediation, and management, we uncovered recurring friction points that shaped our design direction.

Insights we gained:

  1. Progress isn’t visible

  2. Feedback doesn't flow

  3. One system can’t fit all, but one interface must serve all

Solution

Design Decisions

Instead of designing separate tools for each role, I structured the system around shared decision points across the review lifecycle, allowing feature reuse while preserving role-specific focus and accountability.

While the system includes multiple supporting flows, the primary design focus was placed on decision-heavy moments where attention, context, and accountability most directly impact outcomes.


The following iterations show how these decision points were refined through concrete design trade-offs.

Solution

Iteration snap

Building on the mapped-out cross-role journeys and shared feature needs, I explored multiple interaction patterns to balance usability, development effort, and business constraints.

These early decisions were made through quick ideation and close alignment with business and engineering, which helped reduce risk before moving into detailed design.

Solution

Final design highlights

Each of these design highlights responded to the challenges uncovered during early research and workflow analysis, helping resolve key pain points around clarity, task ownership, and team coordination.

Case Allocation with Dashboard

Helped managers make faster and more informed allocation decisions

by combining real-time team availability, case load, and status overview into a single screen.


Document/Audio Case Check

Reduced context switching and improved efficiency

by tailoring the case check interface to match different workflows in branches and contact centres, including inline audio playback and transcript access.


Remediation Query

Streamlined cross-role communication

by redesigning the query workflow into a collapsible side panel, enabling clear categorization, quicker reference, and shared visibility of case history and discussions.


Design System

The Side Panel

I set the design rules for the new side panel component including size and spacing

Reflection

Takeaways

This was the first end-to-end product I led from early research to design execution, marking the beginning of my professional journey.


This project reminded me that thoughtful design isn’t always about reinventing interfaces, it’s about aligning with real-world constraints while still improving clarity, efficiency, and trust.


By grounding early ideation in user needs and system limitations, I was able to facilitate alignment across design, engineering, and business.


Balancing usability with implementation effort became a recurring challenge, but also a valuable opportunity to practice strategic prioritization, advocate for scalable patterns, and build shared understanding among teams.

Role

UX/UI Designer