Voice AI based Contact Centre Solution

Business Problem Statement

• Manual identification of service request type from a live call conversation is a cumbersome process prone to inaccuracies and misrepresentation.
• Manual data entry and form completion leads to longer call duration and unsatisfied customer feedback. This can also lead to unnecessary call drops or callbacks.
• Most contact centre solution cannot detect knowledge articles in an automated manner leading to a CSR manually looking for document and articles while putting a call on hold.
• Cross Sell/Up sell opportunities are also limited due to reliance on a CSRs capability to identify, remember and pitch the right products.
• Call urgency identification is also a manual process dependent on the CSR’s ability to identify the customer sentiment.

Solution Benefits

• Capturing Request Category by context making the identification of service type automated and accurate.
• Reduced call handling time because of automated request type identification, auto filling of fields, speech guidance and real time identification of relevant documents.
• Conveying targeted messages based on adaptive analysis of conversation.
• Targeted product suggestion/upsell leveraging Pega CDH capability to provide the most relevant products and offers.
• Targeted performance management by tracking script adherence, call timing and customer sentiment.
  

Solution Overview

The Voice AI Contact Centre solution assists CSRs during live phone interactions by providing contextually relevant real-time suggestions, such as case and knowledge suggestions, form autofill suggestions, and script adherence notifications, in order to reduce average handle time, reduce callbacks, improve compliance, boost agent productivity, and accelerate new CSR onboarding. It also provides recommendations on appropriate cross sell/up sell products based on the context of the conversation. On conclusion of the call, the solution provides extensive insights on the customer sentiment, competitor analysis as well as agent performance parameters. The solution is powered by real-time speech-to-text transcription and natural language processing (NLP).

• Inbuilt Voice to text transcription capability for real time speech capture and transcription facility.
• Real time adaptive analysis of conversation and recommendation of appropriate task categories and auto population of associated forms and fields on screen.
• Recommendation of most contextual customer communication based on conversation topic and real time measure of the script adherence.
• Built on Pega CDH and Customer Servicing - an extensible, expandable or shrinkable solution as per the organizational process fit.
• NLP Capability for real time analysis of customer sentiment and assign urgency.
• Adaptive analytics for proactive offers.
• Smart Context detection and recommendation of relevant knowledge articles.
• Competitor analysis and agent performance evaluation.

Solution Architecture

Setup the Phone connectivity (Avaya/Amazon Connect/Pega Call) with the Pega CS solution and configure the CTI server. Simultaneously configure a desktop agent in the Pega CS application. Once the setup is completed, the solution is ready for handling calls. During a call, the Pega Voice AI converts the phone audio into text and run analytics on the same. Voice AI detects the context and displays a suggested case in the Interaction Portal. It also auto fills forms and searches through an organization's knowledge base to identify relevant articles. The CDH component inbuilt within the solution parallelly analyses the conversation transcript to identify the right business issues/groups and trigger the applicable actions. This displays the most relevant offers/products to the CSR creating cross sell/up sell opportunities. On conclusion of the call, the NLP module of the solution detects the right customer sentiment to accordingly assign an urgency. The module can also analyse the CSR performance by analysing script adherence, call hold time, competitor references and other parameters.