This guide explains how to enable Conversational Commands for LivePerson Functions.

With Conversational Commands, agents are able to invoke lambdas from the Agent Workspace by typing / into the messaging input in the connections panel. This empowers agents to use a bank of preconfigured, account-specific commands, to enrich the conversation with integrated functions that will help them interact efficiently with the consumer.

Along with the invocation, the function is sent a payload containing the conversation ID and the arguments provided by the agent. This payload can then be used in the function for further processing and referencing. The result of the function has to either be a string or a number and will be shown in the Agent Workspace as a message in the transcript.

Note The response is not actually stored in the conversation transcript, it will only be stored temporarily in the browser and is only visible to the agent.

Permissions

Agents need the following permissions to use Conversational Commands:

• Agent Permission: Use enhanced Agent Workspace and FaaS-Invocation permission has to be enabled.

Conversational Commands events for Function Invocation

Every function is triggerd (or "invoked") via an event. Conversational Commands uses one triggering event for all functions:

Conversational Command

All deployed lambdas with this event assigned to them as their triggering event will be available as Conversational Commands in the Agent Workspace.

Step 1 - Create function

Create a new function using the Conversational Command event as the triggering event. The name of the function will be displayed as the name of the Conversational Command in the Agent Workspace.

The description will also be shown in the Agent Workspace. To give the Agent a hint on how to use the function, we recommend providing the available arguments for the function in square brackets in the description.

Here is an example for the description of a Conversational Command that sends the transcript of the conversation via email:

[email, subject:optional] - Sends the transcript of the conversation to the provided email address.

Step 2 - Edit the Function

Adjust the code in the template according to your needs by modifying the function. On the right side you can see an example of the payload (in the sidebar, which you might need to open).

As mentioned above, the function must return either a string or a number. This response will be displayed in the Agent Workspace as a message. If the function returns anything else, the Agent will only be shown a generic success message. If the lambda returns an Error, this will be presented to the Agent as an error message.

Step 3 - Deploy the function

Just like any other function, this function must be deployed before it can be used. Please see this document for more information on how to deploy your function. At this point, you can also test your function.

Payload details