This customer is one of the world’s largest banking and financial services organizations. It serves its customers from offices and branches in cities across the United States including New York, Los Angeles, Chicago, Atlanta, Houston, Seattle, Miami, San Francisco, and Washington, DC. Through its Wealth and Personal Banking, Commercial Banking, Private Banking, Global Banking, and Markets and Securities Services, the bank offers a full range of traditional banking products and services to individuals, including high net worth individuals, small businesses, corporations, institutions and governments.
The bank's operations team in the United States handles over 36 million calls per year, many of these via their existing Interactive Voice Response (IVR) system. Our technology was implemented on the bank’s voice self-service retail banking voice application. This application was designed primarily for handling U.S. customer inquiries on personal checking accounts, though it also supported home loans, car loans, student loans, and similar retail services.
The client’s voice application handles inquiries for available balance, funds transfer, transaction information, payment history, and general information. It serves account holders via publicly accessible toll-free numbers. While the IVR solution front-ended all incoming phone calls, many callers were opting for an agent and the average handle time for automated calls was high.
During the initial proof of concept (PoC) period with the customer, we implemented our technology to determine what effect dynamically adjusting the audio playback rate of voice prompts in their IVR would have on voice self-service performance. During this period, we ran A/B tests on over 40,000 phone calls over a two week period. We also used Gyst Analytics to collect data on caller behavior as it relates to engagement within the voice application. Existing voice prompts were speed adjusted in direct relation to individual caller skills.
During the initial proof of concept (PoC) period with the customer, audio playback speed adjustment levels of 102, 115, 117, 119, and 121 percent were used. A playback level of 100 indicates the normal playback rate of the audio, 115 represents 115 percent of normal, and so forth.
In summary, the results indicated:
After deploying our technology, the customer realized a decrease in the average handle time in the voice self-service system from 90 seconds down to 83 seconds. Simultaneously, they experienced an improvement in the self-service rate within the voice system - fewer calls went to agents as a result. additionally, callers using the speed adjusted audio had 27.22% more engagement (Conversation Turns) with the IVR over callers using standard audio. These callers also encountered 18.5% fewer error messages and thus, had to reenter information 18.5% fewer times.
The bottom line was a five month ROI on the technology for the bank.
