Enable Voice ID - Amazon Connect

Enable Voice ID

There are two ways you can enable Voice ID for your instance:

Before you begin

Before you get started, complete the following tasks.

Grant the required permissions

You must grant the required permissions to IAM users, groups, or roles. For more information, see AmazonConnectVoiceIDFullAccess.

Decide how to name your Voice ID domain

When you enable Voice ID, you are prompted to provide a friendly domain name that's meaningful to you such as your organization name, for example, Voice ID-ExampleCorp.

Create an AWS KMS key to encrypt data stored in the domain

When you enable Voice ID, you are prompted to create or provide an AWS KMS key. It encrypts the customer data stored by Voice ID such as audio files, voiceprints, and the speaker identifiers.

Step-by-step instructions for creating these KMS keys are provided in Step 1: Create a new Voice ID domain and encryption key.

Data at rest—specifically, freeform fields that you provide plus audio files/voiceprints—are encrypted under the KMS key you choose. Your customer managed key is created, owned, and managed by you. You have full control over the KMS key (AWS KMS charges apply).

When making calls to Voice ID for anything other than CreateDomain or UpdateDomain, the user making the call requires kms:Decrypt permissions for the key associated with the domain. When making calls to CreateDomain or UpdateDomain, the user also requires kms:DescribeKey and kms:CreateGrant permissions for the key. When you create (or update) a Voice ID domain, it creates a grant on the KMS key so that it can be used by Voice ID asynchronous processes (such as speaker enrollment) and by the Amazon Connect service-linked role during your contact flows. This grant includes an encryption context specifying the domain with which the key is associated. For more on grants, see Using grants in the AWS Key Management Service Developer Guide.

If you create a domain and associate it with one key, store some data, and then change the KMS key to a different key, the old data is still stored with the old key. Voice ID doesn't re-encrypt the old data. This means in certain scenarios, the user would require access to the previous KMS key associated with the domain.

Tip

You can create KMS keys or provide an existing KMS key programmatically. For more information, see Amazon Connect Voice ID APIs.

Step 1: Create a new Voice ID domain and encryption key

Following are instructions for how to create a new domain and encryption key.

  1. Open the Amazon Connect console at https://console.aws.amazon.com/connect/.

  2. On the instances page, choose the instance alias. The instance alias is also your instance name, which appears in your Amazon Connect URL.

  3. In the navigation pane, choose Voice ID. Read the BIPA Consent Acknowledgement, and accept if you agree.

    This is a requirement to enable Voice ID and is needed only once per account, across all Regions. This cannot be done using APIs.

    For more information about the Biometric Privacy Act (BIPA) in general, see this Wikipedia article.

  4. In the Domain setup section, choose Create a new domain.

  5. In the Domain name box, enter a friendly name that's meaningful to you, such as your organization name, for example, VoiceID-ExampleCorp.

  6. Under Encryption, create or enter your own AWS KMS key for encrypting your Voice ID domain. Following are the steps to create your KMS key key:

    1. Choose Create KMS key.

    2. A new tab in your browser opens for the Key Management Service (KMS) console. On the Configure key page, choose Symmetric, and then choose Next.

    3. On the Add labels page, add a name and description for the KMS key, and then choose Next.

    4. On the Define key administrative permissions page, choose Next.

    5. On the Define key usage permissions page, choose Next.

    6. On the Review and edit key policy page, choose Finish.

      In the following example, the name of the KMS key starts with bcb6fdd:

    7. Return to the tab in your browser for the Amazon Connect console, Voice ID page. Click or tap in the AWS KMS key for the key you created to appear in a dropdown list. Choose the key you created.

  7. Choose Enable Voice ID.

You've enabled Voice ID for your instance. Next, in Step 2 you configure how you want Voice ID to work in your contact flow.

Step 2: Configure Voice ID in your contact flow

In this step you add the required blocks to your contact flow and configure how you want Voice ID to work.

  • Play prompt: Add this block before the Set Voice ID block to stream audio properly. You can edit it to include a simple message such as "Welcome."

  • Set Voice ID: After the Play prompt block, add the Set Voice ID block. It should be at the start of a call. You use this block to start streaming audio to Amazon Connect Voice ID to verify the caller's identity, as soon as the call is connected to a contact flow. In this block you configure the authentication threshold, response time, and fraud threshold.

  • Set contact attributes: Use to pass the CustomerId attribute to Voice ID. The CustomerId may be a customer number from your CRM, for example. You can create a Lambda function to pull the unique customer ID of the caller from your CRM system. Voice ID uses this attribute as the CustomerSpeakerId for the caller.

  • Check Voice ID: Use to check the response from Voice ID for enrollment status, voice authentication, and fraud detection, and then branch based on one of the returned statuses.

Example Voice ID contact flow

Caller not enrolled

  1. When a customer calls for the first time, their CustomerId is passed to Voice ID using the Set contact attributes block.

  2. Voice ID looks for CustomerId in its database. Since it's not there, it sends a Not enrolled result message. The Check Voice ID block branches based on this result, and you can decide what the next step should be. For example, you might want agents to enroll the customer in voice authentication.

  3. Voice ID starts listening to the customer's speech after the contact has encountered the Set Voice ID block, where Voice ID is enabled. It listens until it accummulates 30 seconds of net speech or the call ends, whichever happens first.

Caller enrolled

  1. The next time the customer calls, Voice ID finds their CustomerId in the database.

  2. Voice ID starts listening to the audio to create a voiceprint. The voiceprint that is created this time is used for authentication purposes so Voice ID can compare if the caller had been enrolled previously.

  3. It compares the caller's current voiceprint with the stored voiceprint associated with the claimed identity. It returns a result based on the Authentication threshold property you configured in the Set Voice ID block.

  4. After it evaluates the speech, it returns the message Authenticated if the voiceprints are similar. Or it returns one of the other statuses.

  5. The contact is then routed down the appropriate branch by the Check Voice ID block.