success story

Virtusa and leading US bank transform contact center with conversational AI

The Challenge

Our client, one of the largest global and Fortune 100 banks, was running a computer-assisted instruction (CAI) program in their consumer banking platform. However, the containment rate of the chatbot for their main branded cards was plateauing at just 18% for over 1.5 years. Furthermore, contact center operating costs were very high, and the computer-adaptive testing (CAT) program was ineffective in containing call volumes. Prices for chat leakage to an agent ranged from $2.50-$3.00, while cost center calls ran $3.50-$5.00 per call.

The Solution

Virtusa holds extensive expertise in delivering large language models (LLM) and natural language processing (NLP) solutions, and we built several tools to accelerate NLP optimization as part of an innovation as a service model for our client.

We built and deployed the following solutions:

  • NLP tools to augment human reviewing and data tagging with an embedded unsupervised learning algorithm, enabling human reviewers to tag 5x the amount of data needed for model training. 
  • A built-from-scratch mapping algorithm to enable production monitoring for accuracy and intent capture rate. 
  • A built for CAI text clustering engine to identify the distribution of users’ objectives and enable teams to identify intents without manual effort.
  • An end-to-end voice-enabled bot to automate interactive voice response (IVR) with integration into Google ASR.
Solution: Virtusa and leading US bank transform contact center with conversational AI
The Benefit

Post-solution deployment, the client experienced a significant increase in the primary bot containment rate and saved more than $10 million annually per bot. Virtusa delivered a proof of concept (POC) with a customized LLM (GPT) model integrated into the kore.ai platform to enable intent recognition. Moreover, the POC initiated a discussion at the CIO level with client leadership to explore considering active investments in LLMs to boost the CAI program further.

Other benefits of the AI-induced contact center transformation include:  

  • The client’s primary bot containment rate increased from 18% to 30% in just three months
  • Current bot containment rate at 60% 
  • More than $10 million saved annually per bot
  • Implementation of intelligent search for mobile apps and online platforms to enable effective digital self-service
  • Nearly 100,000 searches conducted daily to reduce the contact center load

 

Virtusa's Contact Center Transformation

Unlock superior customer experiences

Related content