AWS Partner Network (APN) Blog

Transforming Customer Experience and Boosting Retention with AI-Powered Contact Centers

By Shivansh Singh, Sr. Partner Solutions Architect, I&A – AWS
By Arijit Majumder, Global Alliance & Marketing Manager – Quantiphi

Connect with Quantiphi-1

Today’s global marketplace relies heavily on contact centers for streamlining, maintaining, and maximizing customer service and sales at scale.

However, with the increase in data at the hands of rapidly growing enterprises, contact center executives are faced with the daunting task of combing through vast catacombs of unstructured data within a limited time.

Moreover, the pandemic has exposed a massive imbalance in the supply chains of industries like healthcare. With the increasing gap between demand and supply, healthcare providers are facing a surge in their contact center queries. They require innovative solutions to augment contact center agents to cope with the chasm they’re now faced with.

In this post, we’ll analyze the role of machine learning (ML) solutions in transforming contact centers. We will highlight the key aspects of Quantiphi’s contact center intelligence (CCI) solution built on Amazon Web Services (AWS), and describe how it helped a U.S.-based consumer healthcare organization address contact center challenges by using custom artificial intelligence (AI) and ML techniques.

Quantiphi is an AWS Advanced Consulting Partner with Competencies in Machine Learning, Data and Analytics, Migration, and DevOps. Quantiphi is also a launch partner for AWS Contact Center Intelligence, owing to years of experience building next-generation contact centers driven with cutting-edge AI technology.

Challenges in Contact Centers: An Overview

The traditional customer journey in a contact center entails an interaction with a live agent. However, the drawbacks of this approach are multifold: operational hours for contact centers are restricted to eight to nine hours per day, with no way of reaching the contact centers at irregular times.

Moreover, during customer-agent conversations, there is often a loss of context on the part of the agent. Lack of a structured knowledge base can cause agents to spend most of the call working to understand the query. This can lead to an increase in the time spent on call and negatively affect customer satisfaction (CSAT) and the contact center’s Net Promoter Scores (NPS).

Thus, contact centers need to be armed with next-gen solutions—live agent support, real-time transcription, intent identification, sentiment analysis, identification of key performance indicators (KPIs), customer relationship management (CRM), and insight generation—to help agents address customer queries better and faster.

Quantiphi offers a CCI solution suite that combines state-of-the-art AI/ML services from AWS with custom solution models to modernize contact centers and help organizations enhance their customer interactions.

Quantiphi’s AI-Powered Contact Center Solution

A typical Contact Center Intelligence solution is made up of three components: self-service virtual agent, agent assist, and insight generation dashboard for post-call analytics.

Self Service Virtual Agent/Virtual Assistant

The self-service virtual agent or virtual assistant is either a chatbot or voicebot with omnichannel self-service capabilities, and it’s the first point of interaction for the customer.

With extensive intent identification models running on the backend, this AI-powered customer assistant is equipped to answer standard and complex queries around the clock with an intelligent fallback mechanism. If the query gets too complex, the call is forwarded to the live agent with appropriate context to avoid repetition for the customer.

The solution uses Amazon Lex, a conversational AI service for building conversational interfaces into any application using voice and text. It offers speech-to-text conversions and intent understanding capabilities for building applications with highly engaging user experiences and lifelike conversational interactions.

Real-Time Analytics and Agent Assist

The live agent will have the caller’s conversation history to get relevant information about the caller. This information is provided by an AI-powered agent assist, which can pull up documents from the knowledge repository for the live agent by identifying intents and relevant keywords spoken by the caller.

The agent assistant uses an intelligent Natural Language Processing (NLP) model to analyze the conversations in real-time and provide suggestions. The AWS solution suite uses Amazon Transcribe for adding speech-to-text capabilities into applications. It automatically adds speaker diarization, punctuation, and formatting so the output closely matches the quality of manual transcription at a fraction of the time and expense.

This service is also capable of maintaining customer privacy. When instructed, Amazon Transcribe can identify and redact sensitive personally identifiable information (PII) from the supported language transcripts, allowing contact centers to review and share the transcripts for customer experience insight and agent training.

Post Call Analytics

After call termination, an insights dashboard is generated that provides information on customer satisfaction metrics, agent performance analysis, summarization of the call, and trend analysis. This dashboard facilitates monitoring the health of the contact center for optimizing operations, as well as providing better customer and employee experiences by tracking relevant KPIs.

Insight generation is performed using Amazon SageMaker, Amazon Comprehend, and Amazon QuickSight. Amazon SageMaker is used to build and deploy ML models on the cloud, while Amazon Comprehend is used to find insights and relationships in a text.

Amazon QuickSight is an ML-powered business intelligence (BI) tool for building dashboards on the cloud. The advantage of this service is that dashboards can be accessed from any device and seamlessly embedded into applications, portals, and websites.

Quantiphi’s solution accelerators for these three areas can significantly reduce development time and cost.

Let’s dive into how Quantiphi used the accelerators to help an award-winning consumer healthcare company that provides management support services to healthcare institutions.

Customer Use Case: U.S.-Based Consumer Healthcare Company

In the client’s contact centers, the agents were manually summarizing calls, identifying customer information, and searching through records to answer customer queries. This process was error prone and not customer friendly.

To combat these issues, the organization wanted a solution that could listen to the conversation between customers and agents and extract caller information and important keywords in real time to help agents serve the customers’ requests better.

Quantiphi developed an application with an intuitive interface to assist the agents while responding to queries and capturing information without much hassle.

The application allowed agents to validate callers, capture their personal information, and pin relevant keywords said by the customer for the agent to prepare a more informed response. These keywords were then used to pull up documents from the repository for agents to prepare quick responses.

Over the past few months, the solution has provided real-time transcription for over three million minutes of calls per month with an average keyword extraction accuracy of over 85%.

The agents are able to reduce the call duration by using real-time assistance and therefore are able to attend to more callers. The application will be enhanced in the future to provide a one-stop solution for tracking KPIs and monitoring the health of the contact center.

Quantiphi built this robust and scalable solution for real-time transcription by using a combination of AI/ML services provided by AWS.

Figure 1 – Real-time transcription solution architecture.

Figure 1 – Real-time transcription solution architecture.

Let’s dive into the solution architecture for the solution:

  • Amazon Chime Voice Connector is used to receive the call stream between the agent and the caller separated by speakers.
  • Amazon Kinesis Video Streams streams the call into the system, and AWS Lambda is triggered for multiple functions at various stages of the architecture for call transcription, keyword extraction, and post processing.
  • Entity extraction from the transcripted calls occurs using custom logic and Amazon Comprehend. All transcripts, entities, and keywords are stored into tables in Amazon DynamoDB.

AWS Quick Start: Expedite Your Journey with AWS and Quantiphi

Over the past few years, Quantiphi noticed an increasing need from customers across industries to modernize their contact center operations.

The first step to modernization involves making call transcriptions available to the cloud for analytics. To help organizations expedite this journey, Quantiphi built a Quick Start in collaboration with the AWS Integration & Automation team.

The solution serves as a stepping stone for organizations to generate actionable insights from their contact centers and improve operational efficiency. It sets up real-time call Analytics and provides a dashboard to support agents by displaying live call transcripts, keywords, and call metadata.

Check out the Quantiphi Real-Time Call Analytics Quick Start >>


The contact center industry handles a vast amount of information every day. Accurate transcription, storage, and insight generation from the information is complicated, error prone, and time consuming.

By using Quantiphi’s Contact Center Intelligence solution powered by the AWS CCI solution suite, organizations can enhance their customer experience, and ease operations for their agents.


Quantiphi – AWS Partner Spotlight

Quantiphi is an AWS Competency Partner and AI-first digital engineering company driven by the desire to solve transformational problems at the heart of business.

Contact Quantiphi | Partner Overview

*Already worked with Quantiphi? Rate the Partner

*To review an AWS Partner, you must be a customer that has worked with them directly on a project.