AWS Machine Learning Blog

Enhancing customer service experiences using Conversational AI: Power your contact center with Amazon Lex and Genesys Cloud

Customers expect personalized contact center experiences. They want easy access to customer support and quick resolution of their issues. Delighting callers with a quick and easy interaction remains central to the customer experience (CX) strategy for support organizations. Enterprises often deploy omni-channel contact centers so that they can provide simple mechanisms for their customers to access customer support. But even with these efforts, callers face long wait times, especially during peak hours, which can lead to lower CSAT scores. In addition, organizations have to manage support costs as their footprint expands. As the customer base grows, operational costs for managing a contact center can rapidly increase.

With Amazon Lex bots, you can use conversational AI capabilities to provide highly engaging and lifelike conversational experiences. Organizations can use Amazon Lex to automate customer service interactions and deliver faster responses to queries. As a result, customer issues are resolved in real time, reducing wait times and driving higher satisfaction. You can use Amazon Lex to handle the most common problems encountered by customers. Furthermore, complex issues that require human intervention can be seamlessly handed over from the Amazon Lex bot to a human agent. Augmenting your contact center operations with Amazon Lex bots provides an enhanced caller experience, while optimizing your operational costs with self-service automation. In addition, you can seamlessly scale your contact center operations on the AWS Cloud as your user base grows.

We’re excited to announce Amazon Lex V2 bot support on the Genesys Cloud platform. With this launch, you can build an Amazon Lex bot and set up your contact center in minutes.

About Amazon Lex V2 APIs and Genesys Cloud

Amazon Lex launched V2 APIs and a new console interface that makes it easier to build, deploy, and manage conversational experiences. The Lex V2 console and API enhancements provide support for multiple languages in a single bot, enables simplified versioning, and provides builder productivity tools. These features provide you more control over the bot building and deployment processes.

Genesys Cloud (an omni-channel orchestration and customer relationship platform) provides a contact center platform in a public cloud model that enables quick and simple integration of AWS Contact Center Intelligence (AWS CCI) to transform the modern contact center from a cost center into a profit center. As part of AWS CCI, Genesys cloud integrates with Amazon Lex, Amazon Polly (text to speech) and Amazon Kendra (intelligent search) to offer self-service conversational AI capabilities.

Key features

Genesys Cloud uses the continuous streaming capability with Amazon Lex V2 APIs to enable advanced IVR conversations. With this integration, you can now enable the following:

  • Interruptions (“barge-in”) – Callers can now interrupt the bot and answer a question before the prompt is completed
  • Wait and Continue – Callers can instruct the bot to wait if they need time for retrieving additional information during the call (such as a credit card number or booking ID)
  • DTMF support – Callers can provide information via speech or DTMF interchangeably
  • SSML support – You can configure prompts within the Amazon Lex bot using SSML tags, enabling greater control over speech generation from text
  • Configurable timeouts – You can configure how long to wait for the customer to finish speaking before Amazon Lex collects speech input from callers, such as answering a yes/no question, or providing a date or credit card number

Creating the bot

Let’s create a banking bot as an example and integrate with Genesys Cloud for IVR-based interactions. For a step-by-step process to build an Amazon Lex bot, refer to banker bot workshop. You can also download the bot and import it using the Amazon Lex V2 console.

In addition to the intents presented in the workshop, we add a SpeakToAgent intent to enable handing over the conversation to a human agent based on user requests.

Enabling the integrations

The Amazon Lex V2 integration is available for installation via Genesys AppFoundry. You need an active subscription for premium applications to access the Integration page from the Genesys Cloud Admin dashboard. Genesys also offers a free trial for validation purposes.

1. Configure the IAM role

As invocations for Amazon Lex take place in your AWS environment, you configure an AWS Identity and Access Management (IAM) role with proper permission for Genesys Cloud to assume the role and use resources.

  1. Create an IAM role and select trusted entity to be Another AWS account.
  2. Enter the Genesys Cloud production ID 765628985471 in the Account ID field.
  3. As part of the AWS best practices, you should select Require external ID and enter your organization’s ID to prevent the confused deputy problem and enhance integration security.

By default, IAM roles don’t have permission to create or modify AWS resources. For Genesys Cloud to successfully access Amazon Lex bots, a few permissions are required.

  1. Choose Create Policy and enter the following JSON blob into the policy editor.
{
     "Version": "2012-10-17",
     "Statement": [
          {
               "Sid": "GenesysLexPolicy",
               "Effect": "Allow",
               "Action": [
                    "lex:List*",
                    "lex:Describe*",
                    "lex:StartConversation",
             		"lex:Recognize*",
 					"lex:DeleteSession",
             		"lex:GetSession",
  			  		"lex:PutSession"
              		  ],
                "Resource": "*"
          }
     ]
}
  1. Attach the policy to the role created previously.
  2. Copy the role ARN and configure it within Genesys Cloud.
  3. Save and set the integration status to activate the bot.

2. Configure Amazon Polly

To use Amazon Lex for a voice bot, you set up the text to speech (TTS) capability. Genesys Cloud supports several TTS engines, including Amazon Polly. You can install and configure the Amazon Polly integration following the Genesys documentation. You can then select Amazon Polly as the engine and configure the voice you prefer. To keep the IVR voice consistent in the call flow, the Amazon Polly voice selected in Genesys Cloud should be the same voice configured in your Lex bot. For additional details, see a list of available voices and the associated characteristics.

3. Configure the Genesys Cloud Architect flow

Create an Inbound Call Flow in Architect to orchestrate your bot interaction. You add a Reusable Tasks and use Call Lex V2 bot action to bring in the Amazon Lex bot and design various actions in the call flow.

The integration also allows Genesys Cloud to capture the preconfigured slots as Architect variables. These variables can be used outside of the bot for use-cases such as application of business rules. For example, if a customer provides an account ID that matches with the VIP customer segment, the call can be routed to the priority support queue when transferring to an agent.

4. Configure graceful escalation

When the automated solution can’t fulfill a customer’s request, the interaction should be escalated gracefully. This fallback process allows a human agent to take over the interaction for more complex tasks.

You can save key information from the prior exchange (such as intents, slots, and conversation transcripts) into a script to provide historical context to the agent so that conversations can be picked up seamlessly. This prevents customers from wasting valuable time to repeat the information provided previously.

In the following example, the call is transferred to an available Tier 1 support agent when a customer asks for more help or to be connected to an agent. You can also collect additional context from the customer and hand off to either another bot or human based on specialty.

5. Test the integrations

You can use the native soft phone in Genesys Cloud to make calls as you would with a desktop phone and validate the integration. Enter the bot’s name in the Enter Names and Numbers field and choose Call to follow the prompts.

Summary

Enterprises increasingly invest in automated solutions such as IVR and chatbots as a part of their customer service strategy in contact centers. Automation provides highly available support that handles common tasks without the presence of a live agent, while reducing operational cost.

With the adoption of the Amazon Lex V2 APIs, Genesys Cloud provides an overall improved user experience using the continuous streaming architecture, and enables a more natural customer-bot interaction.

This post outlines the key steps to enable the Amazon Lex V2 integration in your Genesys Cloud environment, and should give you a jump start to create and customize your own chatbot initiative. Check out the following resources for additional information:


About the Author

Anubhav Mishra is a Product Manager with AWS. He spends his time understanding customers and designing product experiences to address their business challenges.

 

 

 

Jessica Ho is a Solutions Architect at Amazon Web Services supporting ISV partners who build business applications on AWS. She is passionate about creating differentiated solutions that unlock customers for cloud adoption. Outside of work, she enjoys spoiling her garden into a mini jungle.