Citizen Developers and Pro Developers have access to the AI Skills package in the Bot editor based on licensing and role.

The Automation Admin creates a custom role to assign access to AI Skills to the users based on their role and task. An AI Skill consists of a Model connection (connecting to a foundational model), Prompt (consisting of a fine-tuned prompt), and Prompt Input (values defined in the AI Skill by the Pro Developer when creating it). The prompt input values can be entered directly within the automation.

Users can invoke the Execute action action in their automation by navigating to AI Skills > Execute, and selecting an AI Skill from the available list.
Note: Use Bot Agent 22.60.10 and later to run your bots successfully.

Prerequisite

  • You require the Bot_Creator and Bot_Runner roles to use the AI Skills > Execute action.
  • You must have access to the relevant folders (whether Private or Public) to access the AI Skills.

Procedure

The following steps show you how to use the AI Skills > Execute action in an automation.
  1. Log in to the Control Room and navigate to Automation > Create new or ‘+’ icon and choose Task Bot. You can also edit an existing Task Bot by accessing it from the Automations tab.
  2. In the Bot editor, add the AI Skills > Execute action to the automation.
  3. Next, click the Execute action in the automation flow to display the details panel.
  4. Click Choose and select an AI Skill from the list you have access to, in your repository.
  5. When you select an AI Skill, the Prompt field gets populated with the prompt text from the AI Skill.

    AI Skills package > Execute action details

  6. The AI Skill inputs field also gets enabled and displays the input values defined in the AI Skill. For example, if the AI Skill has two input values defined, then you would see both the input value fields.
    You can define a variable for the AI Skills inputsfield which gives you the ability to parse information to the AI Skill and get a response back from the model. The AI Skill executes the prompt based on the value entered in this field and returns values for the following fields. These are optional fields and can be defined using ‘If’ variables as per your use case:
    1. Save the response to a variable (optional): Displays the response received from the prompt execution. This is the output value.
    2. Error (optional): Captures the error status. You can define an ‘If’ variable based on the Save the response to a variable output field and define the condition required to see an error message.
    3. Success Status (optional): Captures the success status. Like the Error field, you can define an ‘If’ variable based on the Save the response to a variable output field and define the condition required to see a success message.
    Note: Prompt data details could contain PHI, PII or other sensitive data you choose to enter in the Prompt. We recommend being mindful of this when testing and executing a Prompt.
  7. With the introduction of the Grounded by knowledge base and Grounded by data store Model connections we would additionally display fields with the returned citations in a table format.
    The response goes to a SampleString variable and a table variable gets automatically selected to display the citations. The table displays two keys within each citation. One is the content segment displaying the information chunk and the other is the URI, which is a URL to the chunk of information in the knowledge base or data store the model call is referencing.
    Note: When you run the automation, the execution displays messages showing the referenced citations in message boxes.
  8. Save the automation. Then run and test the automation.