You can create custom actions in Genesys Cloud for use with your data actions integrations. Either copy an existing action (static or custom) to use as a template, or follow these steps to create an entirely new action. When you add an action, you can either add contracts and configuration or import them from another data action.

In Genesys Cloud, use a data action to retrieve the values during the bot conversation. Define the input contract and the output contract. The Output Contract in the data action must match the return message that you configure in your route to endpoint. 

To create a data action in Genesys Cloud, complete these steps:

1. In Genesys Cloud, click Admin.
2. Under Integrations, select Actions.
3. Click Menu > IT and Integrations > Data Actions.
4. Create a custom action for integrations.
5. Add contracts. 
6. Add configuration. 
7. Test the action.

For more information, see Manage individual actions for integrations.

Example: A test API endpoint

Example: Genesys Cloud data action Contracts tab

1. From the Architect home page, click or hover over the Flows menu and select Digital Bot Flows.
2. Click Add. The Create Bot Flow dialog box opens.
3. In the Name field, enter a unique name for the flow. 
4. (Optional) In the Description field, add a short sentence or phrase to describe this flow.
5. Click the Default Language list and select the flow’s default supported language.
6. Click the Divisions list and select the division in which to place the flow.
7. Click Create Flow. The flow’s configuration page opens.

To call a data action from an Architect bot flow, follow these steps:

1. From the Architect toolbox, drag a Call Data Action into the task editor, below the Start element.
2. Under Category, select the appropriate integrations category; for example, Web Services Data Actions.
3. Under Data Action, select the scheduler data action that you created in Genesys Cloud Actions.
4. To configure the Success Outputs, do the following:
   • Set up a variable in the slots.startTime field to store the start DateTime values retrieved from the Call Data Action. For example, Flow.availableStartTimes.
   • Set up a variable in the slots.duration field to store the duration value retrieved from the Call Data Action. For example, Flow.appointmentDurations.
5. Define Failure Outputs and Timeout.

Example: Call Data Action

Prepare a list of timeslot objects to use in your flow by converting the data from your data action. 

To read through the retrieved data, use a Loop action:

1. From the Architect toolbox, drag a Loop action into the task editor, below the Call Data Action.
2. Add an index data name such as Flow.loopindex.
3. Set the loop’s max count to the expression, Count(Flow.availableStartTimes). You can choose any of the variables defined in the Success path of your Call Data Action.
   By following this method, you are running the loop once for every item returned by the data action. Later, you can create a list of timeslot objects and store it in the Collection data type.
4. To create a timeslot collection, 
   1. From the Toolbox, expand the Data category and drag an Update Data action into the Loop action.
   2. In the Update Data workspace, click next to Update Statements and select Collections.
   3. With the Collections list open, select Timeslot Collection.
   4. Under Variable Name 1, enter a name for the list that must store the full list of timeslots. For example, Flow.timeslots.
   5. Under Value to Assign 1, add the expression, AddItem(Flow.timeslots, MakeTimeslot(ToDateTime(Flow.availableStartTimes[Flow.loopindex]), ToDuration(Flow.appointmentDurations[Flow.loopindex]))).
      The preceding expression takes the item in the list of Flow.availableStartTimes that matches the current loop index and converts it into a DateTime object. It also takes the corresponding item from the list of Flow.appointmentDurations and converts that into a duration object. Finally, it combines the two into a new timeslot object and adds that object to the Timeslot Collection variable.

Example: Loop action

Example: Update Data action

1. Under Natural Language Understanding, click Slots.
2. Click Slot.
3. Provide a meaningful name for the slot and click Save.
4. Under Natural Language Understanding, click Slot Types.
5. Click the newly added slot type name. 
6. Under Timeslots, Click Add Collection. The Timeslots window appears.
7. Add the expression, Flow.timeslots and click Close.

Example: Timeslots in Expression mode

1. From the Architect toolbox, drag an Ask for Slot action into the task editor, below the Loop action.
2. Under Slot, select the timeslot that you created previously and then complete the relevant fields.
3. Customize the prompt image, title, subtitle, and address as how it must be presented by the bot.
4. Publish the flow.

Example: Ask for slot with time picker customization

Case: Data action fails

If the third-party API encounters an error, the bot flow takes the Failure path of the Call Data Action. Customize the Failure path based on your use case. For example, you can use a Transfer to ACD action to route the interaction to an available agent. 

Case: Data action times out

If the request to the third-party API times out, the bot flow takes the Timeout path of the Call Data Action. The data action may succeed on a second or third attempt. To set up multiple attempts, you can place the Call Data Action within a Loop action with a fairly low Maximum Loop Count. Ensure that you also place an Exit Loop action after the Call Data Action.

Case: Third-party API returns no timeslots

Even if all timeslots are booked, the Call Data Action takes the Success path because the request is processed successfully. The bot may encounter an error as there are no timeslot values available to power the slot. In this case, the message, Error.Expresson.EmptyCollection, is shown. To prevent this error, you can place a Decision block after the Update Data action (inside the existing Loop action). You can use the Decision block to execute the Ask for Slot action only if there are valid timeslot values available. To find if the timeslots list is empty or not, use the expression, IsNotSetOrEmpty(Flow.timeslots).

Case: All timeslots are in the past

Even if all timeslots are in the past, the Call Data Action takes the Success path because the request is processed successfully. However, the bot cannot present a timeslot, which is in the past. To ensure that all timeslots presented to a user are in the future, you can place another Decision block before the Update Data action (inside the Loop action).

The expression, ToDateTime(Flow.availableStartTimes[Flow.loopIndex]) > GetCurrentDateTimeUtc(), helps you to find whether a timeslot is in the past or future date.