--- schema_version: "1" id: 0b7b54f324df81f899fe0a63c87a2e83378f06684f1f4ea5085b0deea7e03012 title: Create the Amazon Lex Bot collection: admin version: current locale: en-us portal_url: https://docs.salesforce.com/admin/current/en-us/ai/264-0-0/agentforce_voice_create_amazon_lex_bot.md canonical_url: https://help.salesforce.com/s/articleView?id=ai.agentforce_voice_create_amazon_lex_bot.htm&release=264.0.0&type=5 source_site: help.salesforce.com content_type: documentation content_hash: 5f0704214c0aa6f3196fab64deeb3afed57e2b36cffb08717d67a0264241f48d description: Create an Amazon Lex V2 bot that sends a voice contact to the Agentforce agent. Keep the bot and the Q in Connect assistant in the same region. products: - Agentforce - Cross Cloud release: "264" --- # Create the Amazon Lex Bot Create an Amazon Lex V2 bot that sends a voice contact to the Agentforce agent. Keep the bot and the Q in Connect assistant in the same region. Create the Lex V2 bot with `AMAZON.QInConnectIntent` pointed at the Q in Connect assistant ARN. Turn on speech-to-speech with Amazon Nova 2 Sonic. Then build the locale, create a version and alias, and associate the alias with the instance. Lex creates the system `AMAZON.FallbackIntent`. 1. Create the bot. Leave error logging off unless the Lex service-linked role can create the log group. If error logging is on and the role doesn’t have that permission, the bot moves to `Failed`. If that happens, delete that bot and create it again. ``` aws lexv2-models create-bot \ --bot-name \ --role-arn \ --data-privacy childDirected=false \ --idle-session-ttl-in-seconds 300 \ --bot-type Bot \ --error-log-settings enabled=false \ --region ``` ![Tip](https://sf-zdocs-cdn-prod.zoominsoftware.com/tdta-ai-generative_ai-264-0-0-production-enus/f8ef6498-d7b2-44e3-9b34-ffd5a850aae9/images/icon_tip.png) Tip For the Lex V2 service-linked role ARN, don’t use AmazonLexTestWorkbenchServiceRole. To get the role, run this command. ``` aws iam list-roles \ --query 'Roles[?starts_with(RoleName, `AWSServiceRoleForLexV2Bots_AmazonConnect_`)].Arn | [0]' \ --output text \ --profile ``` 2. To confirm that the bot is created and available, run this command. ``` aws lexv2-models describe-bot \ --bot-id \ --query '{botName:botName,botStatus:botStatus,failureReasons:failureReasons}' \ --output table \ --profile \ --region ``` 3. Create the `en_US` locale and turn on speech-to-speech. Use a voice that Nova 2 Sonic supports. The example voice is `tiffany`. Use the `amazon.nova-2-sonic-v1:0` model. Lex rejects the `amazon.nova-sonic-v1:0` model. ``` aws lexv2-models create-bot-locale \ --bot-id \ --bot-version DRAFT \ --locale-id en_US \ --nlu-intent-confidence-threshold 0.40 \ --unified-speech-settings '{"speechFoundationModel":{"modelArn":"arn:aws:bedrock:::foundation-model/amazon.nova-2-sonic-v1:0","voiceId":"tiffany"}}' \ --region ``` 4. Add the Q in Connect intent. Don’t add other intents or sample utterances. ``` aws lexv2-models create-intent \ --intent-name QInConnectIntent \ --parent-intent-signature AMAZON.QInConnectIntent \ --bot-id \ --bot-version DRAFT \ --locale-id en_US \ --q-in-connect-intent-configuration '{"qInConnectAssistantConfiguration":{"assistantArn":""}}' \ --region ``` 5. To confirm the status of the bot locale, run this command. ``` aws lexv2-models describe-bot-locale \ --bot-id \ --bot-version DRAFT \ --locale-id en_US \ --query '{status:botLocaleStatus,failureReasons:failureReasons}' \ --output table \ --profile \ --region ``` 6. After the locale is `NotBuilt`, build it. Then wait until the locale is `Built`. ``` aws lexv2-models build-bot-locale \ --bot-id \ --bot-version DRAFT \ --locale-id en_US \ --region ``` 7. Create a bot version. ``` aws lexv2-models list-bot-versions \ --bot-id \ --query "botVersionSummaries[?botVersion=='1'] | [0].{botVersion:botVersion,botStatus:botStatus}" \ --output table \ --profile \ --region ``` 8. Create an enabled alias. ``` aws lexv2-models create-bot-alias \ --bot-alias-name \ --bot-version \ --bot-alias-locale-settings '{"en_US":{"enabled":true}}' \ --bot-id \ --region ``` 9. To get the alias ARN, run this command. ``` LEX_ALIAS_ID=$(aws lexv2-models list-bot-aliases \ --bot-id \ --query "botAliasSummaries[?botAliasName=='scv-third-party-a2a-alias'].botAliasId | [0]" \ --output text \ --profile \ --region ) echo "arn:aws:lex:us-west-2:764189949692:bot-alias/E07YTAR2BZ/${LEX_ALIAS_ID}" ``` The alias ARN is `arn:aws:lex:::bot-alias//`. 10. Confirm the alias is available. ``` aws lexv2-models describe-bot-alias \ --bot-id BOTID \ --bot-alias-id ALIASID \ --query '{name:botAliasName,status:botAliasStatus,version:botVersion}' \ --output table \ --region REGION ``` 11. Associate the alias with the Amazon Connect instance. ``` aws connect associate-bot \ --instance-id \ --lex-v2-bot AliasArn= \ --region ``` The alias ARN is `arn:aws:lex:::bot-alias//`.