OpenAI

OpenAI Integration

Configure an OpenAI integration to power natural-language conversations through your platform channels or a Chatbot Service automation.

Before you begin, create an OpenAI account.

Select the OpenAI card in the pane.
Select Next.

The integration details window has two configuration sections: OpenAI Information for your account settings and Connections for your platform communication channels.

OpenAI Information

In OpenAI Information, configure the following settings from your OpenAI account:

Alias: Enter a custom name for this OpenAI integration.
API key: Enter your API key. If you do not have one, select + Create new secret key on the OpenAI API keys page.

The OpenAI API uses API keys to authenticate requests.

📘
NOTE
Remember that your API key is a secret! Do not share it with others or expose it.

Model: Select the GPT model to use. Models have different capabilities and price points.

GPT-4 is a large multimodal model that accepts text input and returns text output. It can solve difficult problems more accurately than previous models.

GPT-3.5 models understand and generate natural language or code. gpt-3.5-turbo is optimized for chat and also works well for traditional completion tasks.

Models	Description
GPT-4	More capable than any GPT-3.5 model, able to do more complex tasks, and optimized for chat. Will be updated with our latest model iteration.
gpt-4-0314	Snapshot of GPT-4 from March 14th, 2023. Unlike GPT-4, this model will not receive updates and will be deprecated 3 months after a new version is released.
gpt-4-32k	Same capabilities as the base GPT-4 mode but with 4x the context length. Will be updated with our latest model iteration.
gpt-4-32k-0314	Snapshot of gpt-4-32 from March 14th, 2023. Unlike gpt-4-32k, this model will not receive updates and will be deprecated 3 months after a new version is released.
gpt-3.5-turbo	Most capable GPT-3.5 model and optimized for chat at 1/10th the cost of text-davinci-003. Will be updated with our latest model iteration.
gpt-3.5-turbo-0301	Snapshot of gpt-3.5-turbo from March 1st, 2023. Unlike gpt-3.5-turbo, this model will not receive updates and will be deprecated 3 months after a new version is released.

Tone: Select the chat behavior.
- Balanced: Best for common tasks, such as search, with the fastest response time.
- Creative: Best for generating new content; produces longer, more expressive, and slower responses.
- Precise: Prioritizes factual responses and minimizes conjecture.
Context Limit: Set the number of messages the model retains as context. For example, with a limit of 5, the model considers only the five most recent messages. The default is 5 messages.
Max response length: Set the maximum number of tokens the model can return. Tokens are pieces of words; text is split into tokens before the API processes a prompt.
- Short: 50 tokens
- Medium: 100 tokens
- Long: Unlimited tokens
Select Save to save the OpenAI integration settings.

Connections

In Connections, select the platform channels that will use the OpenAI integration. The dropdown menu lists the available channels.

📘
NOTE
If you want to engage (integrate) OpenAI with a Chatbot Service automation, do not add any connections to the OpenAI.
The channels to be used in the OpenAI integration cannot be shared with the platform Keyword or Chatbot type automations as it would create an operational conflict.

Available channel connection types include:

Additional connection types may be available in future releases.

📘
NOTE
Connections may reference senders that are also used in other platform features, such as auto-replies or keyword automations.
Connection conflicts are evaluated during activation, not during configuration. This means that connections can be added successfully even if they later fail to activate.

To configure channels:

Select a connection from the Available Connections dropdown menu. The left side identifies the channel type, and the right side shows the connection code.

Select Add.

Repeat these steps for each channel in the integration. The platform adds each selected channel to the table below the menu.

After you configure OpenAI Information and Connections, select Save.

For a Chatbot Service automation, enter the required OpenAI information, leave Connections empty, and select Save.

The following sections explain how to activate an integration with connections and how to use one with Chatbot Service:

OpenAI Integration with Connections

The saved integration appears as a card on the Integrations tab. The card shows the alias, tone, API key, connection types, and actions to delete, edit, or activate the integration.

Select Start ▶ to activate the integration. The card updates to show its active state, and active connection icons appear in blue.

Select the Details icon to view the channel and sender for each connection. To deactivate the integration, select Stop under Actions.

🚧
WARNING
When activating a OpenAI integration with connections, the platform attempts to activate:

The integration itself

Each configured connection independently

The platform no longer blocks activation before attempting it. Instead, activation outcomes depend on the result of each connection.
The following scenarios may occur.
No connections activated: If none of the configured connections can be activated:

The integration remains inactive

An error message is displayed

No connections are enabled

Partial connection activation: If at least one connection is successfully activated and others fail:

The integration becomes active

A message is displayed indicating that not all connections were activated

Connections that encountered errors remain inactive

Therefore, when an integration is activated with partial connection errors, a snackbar notification is displayed. This message includes a link that opens the Connections Details dialog.
The Connections Details dialog shows:

Each configured connection.

Its activation status, which means what connections are active or inactive.

This dialog is the same one accessible through the Details icon on the integration card.
A common reason for connection failure is when a sender already has an active auto-reply or is being used elsewhere in the platform.

OpenAI integration with Chatbot Service

After you save an OpenAI integration without connections, activate it before associating it with a Chatbot Service automation.

The saved integration appears as a card on the Integrations tab. The card shows the alias, tone, API key, connection types, and actions to delete, edit, or activate the integration.

Select Start ▶ to activate the integration. The card updates to show its active state.

After you activate the OpenAI integration, associate it with your Chatbot Service automation. See Chatbot Service for instructions on configuring an automation that uses the OpenAI integration.

📘
NOTE
The OpenAI integration must be Active to be associated with a Chatbot Service automation.

After you create the Chatbot Service automation and set its status to Live, the integration card displays the Details icon.

Select the Campaigns icon to view the automation associated with the OpenAI integration. The dialog includes the following information:

The dialog displays the integration type and the following details:

Campaign ID: The automation's unique identifier.
Campaign Name: The automation's name.
Integration Date: The date when the linked automation was created and enabled with Live status.

Considerations

You cannot pause or stop an OpenAI integration while it is linked to an automation. If you try, the platform displays a dialog that indicates the action is not permitted.

To remove an OpenAI integration that is associated with automations:

Disable each automation linked to the integration. Open the associated Live automation, select Report, select Pause, and then select Confirm in the confirmation dialog.

Return to Preferences > Integrations and select Pause.

To delete the integration, select the Delete icon under Actions.