Skip to content

Endpoints

Endpoints are connectors between Cognigy virtual agents and external systems or platforms, such as Facebook (Meta) Messenger, Slack, Alexa, and more.

Endpoints support text-based conversations as well as conversations on your contact center phone lines.

Endpoints convert channel-specific input and output, including images, galleries and voice, to the standardized Cognigy Input object. This approach enables the development of platform-independent Flow that require no modification when adding new channels.

Endpoint Types

The platform includes many pre-built Endpoints that can be configured in a couple of steps.

Cognigy Endpoints

The following Endpoints are specific to the Cognigy.AI platform:

Cognigy Endpoint Description
Cognigy Voice Copilot Used for integration between call centers and AI Copilot for voice use cases.
Cognigy Voice Gateway Lets you connect to SIP interfaces, which means that your virtual agent can be connected to the phone network.
Cognigy Webchat An open-source web-based chat widget that can be embedded on any website.

Generic Endpoints

The following Endpoints are provided for custom integrations:

Generic Endpoint Description
REST The REST Endpoint lets you connect to a Cognigy Flow directly through a REST interface.
Socket The Socket Endpoint lets you connect to a Cognigy Flow directly through a socket client.
Webhook The Webhook Endpoint lets you expose Cognigy Flows via Webhook.

Third-party Endpoints

The following Endpoints provide integrations with third-party products:

Third-party Endpoint Description
Amazon Alexa Alexa is Amazon's Virtual Assistant voice AI device. The Endpoint supports all of the devices, such as the Echo, Echo Dot, and Echo Show.
Amazon Lex (Beta) The Amazon Lex Endpoint allows you to connect your Amazon Lex bots with Cognigy.AI. This is achieved by using a Lambda function with your Lex bots.
Avaya CPaaS Using the Avaya CPaaS Endpoint you're able to connect your Cognigy.AI resources to your Avaya CPaaS phone number.
Azure Bot Services The Azure Bot Services Endpoint allows you to connect your virtual agent to all Azure Bot Service channels.
Bandwidth The Bandwidth allows connecting virtual agent to the Bandwidth provider.
Dialogflow DialogFlow is Google's Natural Language Understanding framework. By configuring a DialogFlow Endpoint, you can connect your Flow to all channels that are supported by DialogFlow.
Facebook (Meta) Messenger Messenger is Facebook's messaging app and platform.
Genesys The Genesys Endpoint allows connecting virtual agent to the Genesys Cloud CX platform using a Genesys Bot Connector.
Intercom Intercom is a live chat suite.
LINE Messenger Line is a popular messaging and voice call app.
Microsoft Teams Microsoft Teams is a popular communication and collaboration platform.
RingCentral Engage RingCentral Engage is a customer engagement platform with built-in support for many channels and live-chat capabilities.
Slack Slack is a popular communication and collaboration platform.
Sunshine Conversations Sunshine Conversations is part of Zendesk and helps connect to another set of channels.
Twilio Twilio is a popular cloud communications platform. The Twilio Endpoint lets you connect Flows to Twilio Programmable Voice.
Twilio SMS Twilio is a popular cloud communications platform. The Twilio SMS Endpoint lets you connect Flows to Twilio Programmable SMS.
Userlike Userlike is a popular live-chat platform.
WhatsApp Using the WhatsApp Endpoint integration enables you to connect your Cognigy resources to the WhatsApp Cloud API access via Facebook (Meta) Developer Dashboard.
Workplace by Facebook (Meta) Using the Workplace by Facebook Endpoint integration enables you to connect your Cognigy resources to the Workplace channel.
8x8 The 8x8 Endpoint allows connecting virtual agent to the 8x8 platform.

General Endpoint Settings

Every Endpoint needs to be configured in the following configuration fields:

Snapshot

Endpoints can be pointed to a specific Snapshot if there is at least one Snapshot in the Agent. This enables you to easily deploy different versions of your virtual agent.

While an Endpoint is active, you can replace or remove a Snapshot. Note that active sessions may continue to use the old Snapshot until they time out.

Keep in mind that if you have a configuration such as a Webchat Endpoint using Persistent History, the user's browser will retain message history. Even after the session expires, users will still see old messages, but any new message will initiate a new session.

When a Snapshot is replaced with a new one or is deleted, the following rules will be applied to the active session:

  • If a Snapshot was replaced and not deleted, the active session will continue with the old Snapshot. The reason is that the system cannot recognize if the new Snapshot contains the same Flows and Nodes where the user might be active in their session.
  • If the old Snapshot was deleted, the session will be reset, and the new Snapshot will be used.

For more information, read the Snapshot article.

Override Snapshot Connections

When Agents use a Snapshot, a toggle Override Snapshot Connections will be displayed in the Endpoint editor. Enabling this option allows the Endpoint to override the Connections in a Snapshot with Connections from the project. If a Connection is available only in the Snapshot, it will be used regardless of the toggle setting. If necessary, such Connections need to be manually updated.

Default setting

By default, any Endpoint will continue using the Connections from the Snapshot unless the toggle is turned on.

To override the connection in the designated Agent, begin by either restoring it from a Snapshot or importing Connections from another Agent using a Package. Ensure that the Snapshot selected in the Endpoint Editor is created from the same Agent to ensure a match of Connections.

In this case, if the toggle is enabled, any change in the Connections of the restored Agent will take precedence over the Connections in the selected Snapshot.

Note

You can verify or troubleshoot Connections by their referenceId, which needs to match both in the Project and the Snapshot for the override feature to work. To do that, use the Cognigy.AI API with the Get all Connections request.

Flow

Endpoints need an entry Flow to start the conversation. You can switch Flows during a conversation.

To switch to a Flow during a conversation, do the following:

  1. Go to Deploy > Endpoints and select the Endpoint you want to use, if required.
  2. Open the drop-down menu of the Flow selection field.
  3. Click jump to Flow next to the Flow selection field on the right-hand side. The selected Flow will be displayed and active.

To display the Flow ID of the currently selected Flow, click Manual Input below the Flow drop-down field.

To toggle back to the drop-down selection list of Flows, click the Select from Dropdown link.

Locale

The Localization feature in Cognigy.AI provides tools for enabling multilingual Agents.

You can add a Locale that will be applied across an Agent and allow language-specific output messages to be configured within Flows. The system works with fallback layers, which means that a Locale that has no content configured will fall back to another Locale that has.

The locale selector is only available if you have created at least one additional locale beforehand.

To select a Locale, do the following:

  1. Go to Deploy > Endpoints and select the Endpoint you want to use.
  2. Click expand next to the Locale selection field on the right-hand side to open the drop-down menu.
  3. Select a locale.
  4. Click Save to confirm your selection. The Endpoint is configured to that Locale now if the selected language has been edited.

As a result of the selected locale, your Endpoint, for example, Webchat, will use the selected language in a conversation based on your Flow.

You can test the functionality of your Webchat Endpoint locale setting as follows:

  1. Click Open Demo Webchat at the top right of the Endpoint Editor. The Cognigy Webchat window is displayed.
  2. Enter a text message in the text field at the bottom and click to send the message as a user.
  3. The bot replies as configured in the Flow but using the language of your previously selected Locale.
  4. Check if it works as expected. When no translated content of the selected Locale is defined, a fallback language will be used.

Endpoint URL

The Endpoint URL serves as the interface to the Endpoint. For most channels, you'll need to copy it into the channel's settings to configure the connection to Cognigy.AI.

Enable Endpoint

This setting allows you to quickly enable or disable the Endpoint.

Natural Language Understanding

This setting allows you to select a different NLU engine for the selected Endpoint.

The NLU Connectors that are configured to the agent are available to select from this drop-down menu. Alternatively, the No NLU option can be selected to completely disable the intent and slot mapping for an endpoint.

Data Protection & Analytics

In the Data Protection & Analytics section, you can configure the collection of analytics and choose whether or not to gather any data.

Chatbase

Support of Chatbase tool is ended

Chatbase is an outdated analytics tool and is not supported any longer.

Cognigy.AI can be connected directly to Google's Chatbase analytics platform. When enabled, your Chatbase bot's API key needs to be entered. All relevant data that enters the Endpoint is forwarded to Chatbase.

Dashbot

When enabled, the Dashbot feature will collect analytics data for the selected Endpoint.

Specific Endpoint Settings

You can use the following additional settings for the Endpoint configuration:

More Information