Cloud case studies
Stopping revenue leaks with natural language processing
Wealth management firm implements NLP engine and increases team's capacity
A wealth management firm turned to Deloitte to design and help implement a natural language processing engine to help address revenue leaks.
Our story starts with a large US asset and wealth management firm that was unable to reconcile its sheer volume of paper contracts and invoices. This led to 3 to 4 percent of its revenue leaking out across thousands of transactions due to mismatches between contracts and client invoices. The scary part? Most of the time it went unnoticed. And when someone did notice a discrepancy, the manual reconciling process was both time-consuming and highly prone to human error.
"The manual reconciling process was both time-consuming and highly prone to human error."
Stopping revenue leaks with natural language processing
What happened next
That's where the wealth management firm turned to Deloitte to design and help implement a natural language processing engine to read and understand contracts despite differences in how they were written. The program also included a human in the loop workflow allowing employees to intervene as needed and to provide smarter business insights to decision-makers. Working with Deloitte over ten weeks, the firm identified revenue leakage across 20,000 transactions amounting to 3 to 4 percent of their business. Now it's like one person can do the job of ten, freeing up staff to perform higher-value tasks.
Overall, the team's capacity increased by 50 percent and cost the firm 30 percent less than what it would cost to hire extra team members. The solution also yielded 96 percent accuracy with 1,500 data points entered without error, freeing up resources that were previously spent on rework. The future doesn't belong to workers or machines, but in the power of both working with each other.
By the numbers
Working with Deloitte over ten weeks, the firm identified revenue leakage across 20,000 transactions amounting to 3% to 4% of their business
The solution allowed one person to do the job of ten, freeing up staff to perform higher-value tasks
The team's capacity increased by 50% and cost the firm 30% less than what it would cost to hire extra team members
The solution yielded 96% accuracy with 1,500 data points entered without error
Explore our cloud services and insights
LET'S CONNECT.
