Dropzone for Zendesk Messaging + AI agents: Configuring an AI agent Answer

Follow the below steps if you wish to use the SendSafely Dropzone for Zendesk Messaging in an AI agent workflow. Once the answer is configured, AI agents can insert a button to the Dropzone in Zendesk Messaging and take action after the upload is complete.

Pre-requisite Step

Before setting up an AI Agent Answer, you must first have configured the AWS Lambda Dropzone Connector for use with Zendesk Messaging. Follow the steps outlined in this article.

Step 1: Define an answer that will leverage the Dropzone 

Answers tell an AI agent how to respond to customer questions. In order to use the SendSafely Dropzone, you need to tell the AI agent which answers will require the use of a Dropzone and configure the interaction.

In the Zendesk Admin Center, navigate to the AI agent you want to configure (AI > AI agents > AI agents for Messaging> [AI agent]) and click the "Answers" tab to select the specific Answer that will utilize the Dropzone.

Step 2: Ask for Email

The Dropzone requires a valid email for submission. Therefore, if the customer's email address is not already known by the AI agent you must include a step to ask the user for their email address. After adding Intent and any introductory messages, insert an Ask for details step to collect the Email.

Step 3: Inserting the Dropzone button (API Call)

Next, add a Message directing the user to "Please use the button to securely upload your file(s)", then insert a Make API call step that calls the Dropzone Connector with the following API details:

Request method: POST
Endpoint URL: https://<your-lambda-function-URL>/
Body:
{
"chatBotAction": "insert-extension",
"chatBotConversationEmail": "{{sf.requester.email}}",
"chatBotConversationId": "{{system.sunco.conversationId}}",
"extensionMode": "bot"
}
Authentication: No Authentication
Variables: uploadid $['uploadId']

The Endpoint URL can be either the AWS Lambda Function URL (available from the AWS Function overview screen) or the API Gateway Endpoint (if you have deployed an API Gateway).

Step 4: Dropzone API Call Success/Failure 

• If the API call failed, we recommend adding a Transfer step with "The secure file upload is currently unavailable" message.
• If the API call is successful, add an Options step to allow the customer to either upload more files "I have more to upload", or indicate completion "I have uploaded my file(s)"

Step 5: Insert Link to uploaded files (API Call)

From Step 4 above,

• If the user clicks 'I have more to upload',  Link to another answer and restart the answer flow
• If the user clicks 'I have uploaded my file(s)' , insert a link to the uploaded files using the following steps:

Insert a Make API call step with the following API details:

Request method: POST
Endpoint URL: https://<your-lambda-function-URL>/
Body:
{
"chatBotAction": "insert-link",
"chatBotConversationId": "{{system.sunco.conversationId}}",
"uploadId": "{{uploadid}}"
}
Authentication: No Authentication
Variables: packagelink $['packageLink']

Step 6: Insert Link API Call Success/Failure

• If the API call failed, add a Transfer step with message "We cannot retrieve the package link at this time. Transferring you to an agent."
• If the API call is successful,
  • add a Message stating "Thank you.", and
  • add button with Button link "packagelink" and Button text "Upload Successful" 

Step 6: End flow

At this point the Dropzone flow is complete. You can add a final Transfer step to transfer the user to an agent. When viewing the conversation in the Agent Workspace, the agent will automatically see the uploaded files appear in the SendSafely sidebar app.