loading

Business Impact

  • 0.5s

    Average Response Time

  • 562K

    Interactions handled across 289K sessions within the first 45 days of launch

  • 79%

    Success Rate

Customer Key Facts

  • Location : North America
  • Industry : Banking

Problem Context

In the wake of the COVID-19 outbreak, the customer, one of the largest U.S. financial services companies, began issuing a prepaid debit card to eligible customers at the request of the state agency for providing benefits to citizens. However, their contact centers began experiencing high call volumes with queries related to the benefits card. As a result, a large number of conversations were put on hold or dropped due to long wait times. In addition, there was a lack of consistent responses from their human agents due to the rapidly evolving COVID-19 situation.

Challenges

 

  • Google Sheets as a CMS was not supported in customer’s environment
  • Lack of “Out of Box” customization options for Dialogflow Messenger
  • Intents clashing due to similarity in phrases
Challenges

Technologies Used

Dialogflow
Google Cloud Platform
Node.js
CSV

Implementing An AI-Powered Virtual Agent For Enhanced Customer Service Amid COVID-19

Solution

Quantiphi developed and deployed two virtual agents on two of the client’s websites trained on 100+ custom intents in just six weeks, responding to queries related to client’s retail services, benefits card offering, and stimulus packages provided by the state and federal government amid COVID-19. The virtual agents were also integrated with the client’s internal content management system for easy update and modification of intent responses by their staff and Subject Matter Experts (SMEs). The virtual agents effectively handled over 500 thousand interactions within the first 45 days of launch.

Result

  • Reduced live agent call volumes by handling 12.5K interactions a day
  • Instant resolution to user queries (0.5 seconds) with 79% success rate

Looking for similar project?

Let's Talk

Get your digital transformation started

Let's Talk