Stopping $1 million revenue leakage
This case study highlights how I identified bottlenecks slowing down the legal review team. The outcome of this research helped in stopping revenue leakage and increasing the firm's revenue by $1 million annually.
​
This case study also showcases how I designed an automated ticket distribution system that successfully mirrored complex human decision-making by partnering closely with stakeholders and end-users.
The Problem
In 2024/25, a Global Private Bank saw a surge in the number of clients investing in alternative assets. This led to an increase in the number of legal documents that needed to be reviewed by the firm’s legal review team.
Due to the limited availability of legal reviewers within the firm, several documents had to be outsourced for review. This led to significant revenue leakage for the firm.
The Goal
The goal of this project was to identify and eliminate process inefficiencies to regain this revenue.
My Role
I led the research to identify the bottlenecks that were impacting the efficiency of the legal review team. I also played a key role in helping the design team define the right solution.
Identifying Bottlenecks
I conducted deep-dive interviews with the legal review team to map their end-to-end workflow. By visualising the journey, I identified 3 primary bottlenecks:​
-
Manual Task Allocation: Team leads spent 30 minutes every morning manually assigning tickets to team members via Excel.
-
UI Complexity: A cumbersome, dropdown-heavy interface slowed down error reporting.
-
Communication Gaps: Clarifying legal comments to clients required repetitive, manual follow-ups for complex cases.
Prioritization a Bottleneck
I presented these findings to Product, Design, and Business stakeholders. To drive consensus, I facilitated a prioritisation workshop where we weighed the frequency vs. the impact of each bottleneck.
​
-
The Decision: We prioritised Manual Task Assignment.
-
The Reason: manual assignment caused a ~30-minute daily every day for the entire 25-person team. The other issues only affected 10–20% of complex cases.
Exploring the Solution Space
With the problem defined, I led the team in exploring a wide spectrum of potential solutions. After evaluating several ideas for feasibility and impact, we narrowed our focus to two distinct strategic directions:
​
-
Direction A - No Automation: This involved building an intuitive software interface that replicated the team’s current Excel-based workflow. While this would still give Team Leads full control over order-assignment, it would only partially eliminate the 30-minute daily delay.
-
Direction B - Complete Automation: This involved building an intelligent engine that automatically distributes orders to reviewers. While this would completely reclaim the lost time for the 25-member team, it would be more complex to build, and Team Leads would have very little control over order assignment.
Choosing the Right Path
To decide between these options, I facilitated focus groups with the Legal Review Team Leads. They were initially hesitant about Complete Automation, fearing the system might assign a "high-complexity" order to a junior reviewer or overload an individual’s capacity.
To address these concerns, I conducted deeper research to understand the logic used by Team Leads for assigning orders to Team Members. I also learnt about the different problems that could arise when assigning orders and how Team Leads resolve them.
Based on these learnings, I worked with the Design and Tech teams to refine the "Complete Automation" model.
Choosing the type of queue
When defining the solution, one of the decisions we had to make was choosing between 2 types of queuing models:
-
Multi-channel queue: Each reviewer has a finite number of orders assigned to them based on their skills & capacity.
-
A Single-channel queue: Each reviewer receives one ticket from a common queue to start with. When they complete the review, the next ticket gets assigned based on their skill.
To help with this decision, I facilitated another focus group with the legal review team leads. During this session, I presented visual representations of both options to the team leads to gather their feedback.
The leads overwhelmingly preferred the Single-channel Queue because it automatically accounted for the varying complexity of legal documents. If one order took longer to review, the system simply waited until the reviewer was ready for the next, preventing burnout while ensuring 100% efficiency.
The final solution
I worked closely with Design and Tech teams to build the automated ticket-assignment engine. This improved the current experience in 2 key ways:
-
For Reviewers: They now begin work the moment they log in. They don’t have to wait for the team lead to assign tasks.
-
For Leads: They don’t have to spend 30 minutes every day on Excel assigning tasks. They only need to intervene for special cases like priority overrides.
The Results
-
Reclaimed Time: Eliminated the 30-minute daily "dead time" for all 25 team members.
-
Increased Throughput: Each team member now completes 2 additional reviews per day.
-
Bottom Line: This reclaimed capacity allowed the firm to move reviews in-house, generating an estimated $1 million in additional annual revenue.
