Integrating PolyAI voice agents with Amazon Connect enhances your contact center by automating call handling, securely managing data, and ensuring seamless live-agent handoffs. This guide walks you through setting up and optimizing the integration for efficient customer interactions.

Overview of capabilities

Automated call handling

PolyAI voice agents handle incoming calls routed via Amazon Connect, capturing customer intent and responding with knowledge-based answers. This minimizes the load on live agents while maintaining high-quality service.

Secure data access

PolyAI leverages AWS STS (Security Token Service) to securely retrieve, update, and store call attributes in Amazon DynamoDB. This ensures secure and real-time data exchange during the call lifecycle.

Dynamic routing and workflows

PolyAI integrates with Amazon Connect Contact Flows to support advanced call routing based on customer input, predefined rules, or real-time conditions.

Live-agent handoffs

Seamlessly transfer calls requiring human intervention to Amazon Connect agents with full contextual data, ensuring continuity and personalized customer service.

Real-time analytics

Track call performance, resolution rates, and agent interaction efficiency through Amazon Connect’s analytics dashboard, combined with PolyAI’s detailed call logs.

Additional capabilities

Scalability

Amazon Connect’s cloud-based CCaaS scalability, paired with PolyAI’s AI assistance, ensures calls are handled efficiently even during traffic surges. This guarantees operational continuity and exceptional customer experiences.

Personalization

PolyAI retrieves customer profiles from CRMs and Amazon Connect before calls begin, enabling personalized and contextual interactions.

Insights and business intelligence

With PolyAI’s APIs, businesses can access conversation data and request insights into customer satisfaction and service performance, empowering data-driven decision-making.

Seamless feedback loop

Data collected by PolyAI informs Amazon Connect’s Contact Flows, ensuring calls are directed to the most appropriate flow for optimal outcomes.

How the integration works

  1. Call initiation: Amazon Connect routes inbound calls to a PolyAI-provisioned phone number.
  2. Assistant connection: The call is transferred to the PolyAI conversational assistant, which retrieves contact attributes from Amazon DynamoDB via AWS STS.
  3. Caller interaction: The conversational assistant greets the caller, verifies identity, collects information, and answers queries.
  4. Call resolution: The assistant resolves the query or hands the call back to Amazon Connect for routing to a live agent.
  5. Data update: Information collected by the assistant is pushed back to Amazon Connect for logging and analysis.

Prerequisites

  1. Amazon Connect instance: Ensure you have an active Amazon Connect instance with administrative access.
  2. PolyAI project: Configure your PolyAI project to align with Amazon Connect workflows.
  3. AWS services setup:
  4. Integration credentials: Obtain API keys and IAM roles for connecting Amazon Connect with PolyAI.

Setting up the integration

Configure Amazon Connect instance

  1. Log in to your AWS Management Console and open Amazon Connect.
  2. Navigate to the Contact Flows section and create or edit a flow.
  3. Add the following key components to your flow:
    • PolyAI handoff node: Routes calls to PolyAI voice agents.
    • Lambda function node: Configures a Lambda function for processing call events.
  4. Save and publish the flow.

Configure AWS Lambda function

  1. Create a Lambda function in the AWS Management Console.
  2. Use the provided template for processing call events and securely retrieving attributes from DynamoDB.
  3. Grant the Lambda function necessary permissions using an IAM role with DynamoDB and STS access.
  4. Deploy and link the Lambda function to your Amazon Connect flow.

Set up DynamoDB for call attributes

  1. Create a DynamoDB table to store call attributes.
  2. Define primary keys to manage data, such as CallID or CustomerID.
  3. Link the DynamoDB table to your Lambda function for real-time data updates during call interactions.

Additional onboarding requirements

  1. Provide PolyAI with your AWS account ID and Amazon Connect region.
  2. Create an IAM role named polyai-cf-deployer with the specified policy and trust relationship.
  3. Work with the PolyAI team to deploy the CloudFormation stack, which will create a DynamoDB table, a Lambda function, and an IAM role.
  4. Update your Amazon Connect Contact Flow to route calls to PolyAI bots.

Testing and deployment

  1. Conduct end-to-end tests: Verify that Amazon Connect, Lambda, DynamoDB, and PolyAI components function as expected.
  2. Monitor performance: Use Amazon Connect’s analytics dashboard to track call volume, routing efficiency, and agent handoffs.
  3. Optimize workflows: Adjust contact flows, API configurations, and Lambda function settings based on test results.