Livechat Configuration Tutorial

Create PAT Authentication in the Livechat Developer Console

Log in to the Livechat Developer Console using the LiveChat Agent account that will be taken over by the AI Agent, and follow these steps:

• Go to the settings page, select Personal Access Tokens, and click "Create new token +"

• Create PAT, select "select all" for access scopes

• Copy and save the "Account ID" and "Token", which will be needed later for configuration in "GPTBots - Agent Integration - Livechat".

Build an app in Livechat Build app for receiving and sending messages on the livechat platform.

• Go to the "Apps" page and click the "Build app" button to start creating the app

• Enter the app name, select livechat for Product, and click "Create app" to successfully create it

• Click the app to enter the details page, select the Blocks page, and click the "Add Building Blocks" button

• In the "Add Building Blocks" popup, select the "App Authorization" option and click "Add block"

• Select the Server-side app option and click the "Save" button to successfully create it

• Copy and save the "Client Id" and "Client Secret", which will be needed later for configuration in "GPTBots - Agent Integration - Livechat".

Configure Livechat Integration in GPTBots

• On the "GPTBots - Agent Settings - Integration - Livechat" page, configure the saved information and click the "Submit" button to complete the configuration.

  • AgentEmail: Email address of the LiveChat Agent account that will be taken over by the AI Agent. This agent must be a human customer service account in the same project.
  • Account ID: The Account ID obtained when creating PAT in Livechat (for the LiveChat Agent account being taken over by the AI Agent).
  • Token: The Token obtained when creating PAT in Livechat (for the LiveChat Agent account being taken over by the AI Agent).
  • Client Id: The Client Id obtained when creating the app in Livechat.
  • Client Secret: The Client Secret obtained when creating the app in Livechat.

• Secret key: Enter a custom key of your choice and save it for webhook authentication.

• After completing the "Livechat Integration" configuration in GPTBots:

  • The AI Agent will take over the user conversation of the AgentEmail account, directly responding to user questions.
  • Based on the AI Agent's capabilities, it can support multi-modal information input such as text, images, and files from users.
  • The platform will generate a "Webhooks - Message and Event Reception URL" address, which can be used to create a Webhook in the Livechat backend.## Configure Chat webhooks for advanced features like chat transfer/closure

Create "incoming_event" and "incoming_chat" event

• Return to the Blocks page, click the "Add Building Blocks" button to create webhooks

• In the popup, select the "chat webhooks" option (note that it is not the app webhooks) and click the "Add block" option

• Copy the "GPTBots Webhook URL" address in "GPTBots - Agent Configuration - Integration - Livechat"

• Configure the following options in "Creating Chat Webhook" and click "Save" to create successfully

  • Webhook URL: Fill in the copied "GPTBots Webhook URL" address in the input box.
  • Secret key: Enter the custom key saved earlier
  • Type: Select the "License" option
  • Webhook triggers: Select the "incoming_event" and "incoming_chat" event and check "chat_presence_user_ids"

Install app and verify

• Go to the "Private Installation" page and click the "Install app" button.

• Finally, go to https://my.livechatinc.com/apps/settings/manage-apps/private to check if the installation is complete. If there is an install mark, it means the installation was successful.

Configure Livechat Agent

Select "Team" - "Group" to create a new group. Add the LiveChat Agent account that will be taken over by the AI Agent to this group. Also add a LiveChat Agent account that will be used by human customer service.

• Set the LiveChat Agent account that will be taken over by the AI Agent as: Primary agent.
• Set the LiveChat Agent account used by human customer service as: Backup agent.

Advanced features of Livechat integration

Setting TAG

Once the authentication configuration for Livechat is completed, the "Setting TAG" feature can be enabled. The specific configuration is as follows:

• Select trigger scenarios: The system allows the "Setting TAG" operation to be triggered by two events: "Closing Conversation" and "Transferring Chat."
  • Closing Conversation: When the user actively closes the conversation in the Livechat chat window.
  • Transferring Chat: When the conversation in the Livechat chat window is transferred to another customer service account.
• TAG: The system will automatically retrieve the list of TAGs already set in the Livechat project. When the selected trigger scenario occurs, the AI Agent will choose the most appropriate TAG based on the conversation record and automatically call the "Livechat TAG Setting API" to assign a TAG to the conversation. This aids customer service in swiftly identifying user needs and enhances operational efficiency.

Replying Status

Once the authentication configuration for Livechat is completed, the "Replying Status" feature can be enabled. After enabling this feature, when a user sends a message in the Livechat chat window, they will see a "replying animation," which improves the overall user experience.

Transferring Chat

After the AI Agent takes over the Livechat customer service account, while the AI Agent resolves most customer issues, some scenarios still require human intervention to address specific problems. To support this, GPTBots provides the "Transferring Chat" feature.

Agent Configuration for Transferring Chat

livechat widget to livechat: When users send messages in the Livechat widget chat window, the AI Agent automatically responds. If human assistance is needed, human customer service representatives will respond through Livechat. Users remain in the Livechat widget chat window throughout the entire process. other to livechat: When users communicate through GPTBots widget chat window, WhatsApp, API, Telegram, or other channels, the AI Agent automatically responds. If human assistance is needed, human customer service representatives will respond through Livechat. Users remain in their original channel throughout the entire process.

• Enable the "Human Service" feature in a standard Agent, and select "Livechat - livechat to livechat." If authentication information has been configured in the Agent integration, it will display as Integrated; otherwise, it will display as Not Integrated.
• Transferring Customer Service Group: After completing the authentication configuration, the system will automatically retrieve the names of Agent groups already set in the Livechat project. You can customize the selection of the target customer service group.
• Trigger Conditions: Based on the actual business needs of the enterprise, use natural language to specify the conditions for initiating the chat transfer. For example: When the customer's inquiry contains "return," initiate the chat transfer. When the user is emotionally agitated and the inquiry contains "complaint," initiate the chat transfer.

Flow-Agent Configuration for Transferring Chat

• Enable the "Human Service" feature in Flow-Agent, and add a "Human Service" node to the workflow canvas. If authentication information has been configured in the Agent integration, it will display as Integrated; otherwise, it will display as Not Integrated.
• Transferring Customer Service Group: After completing the authentication configuration, the system will automatically retrieve the names of Agent groups already set in the Livechat project. You can customize the selection of the target customer service group.
• Trigger Conditions: The trigger for Flow-Agent depends on the workflow's running status; therefore, custom trigger conditions are unavailable.

Transferring Chat Effect

When a user sends a message in the Livechat chat window, the AI Agent identifies the best-matched Agent group based on the conversation record and Agent group names, and routes the user's conversation to that Agent group (human support group). A human customer service account in that group will then handle the user's conversation.