Co-Pilot for Automators

Co-Pilot for Automators is an on-demand, intelligent automation assistant. Developers can accelerate building automation by describing goals in natural language for Co-Pilot for Automators to create and update their automation.

Co-Pilot for Automators is embedded directly within the Automation Workspace. The user can build enterprise-level automation with an interactive instruction-based conversation using Co-Pilot. It is non-intrusive, allowing developers to continue building or editing automation without interference while helping accelerate building automation. This provides users with the flexibility to decide when and how to use this feature.

Co-Pilot for Automators follows these principles to ensure maximum efficiency and ease-of-use:
  • embedded conversational experience for building automation with natural language
  • ensures low-learning curve for quick adoption
  • built-in guardrails with enterprise-tuned AI models that only allow Automation 360 tasks

Leveraging capabilities

Developers can overcome common challenges and expedite development by leveraging Co-Pilot for Automators when and where support is most impactful. Co-Pilot for Automators can assist in selecting and sequencing automation tasks and actions to reach development goals. With Co-Pilot for Automators, developers simply need to describe the automation scenario for necessary elements or actions to be added in proper order. Consider the following areas where Co-Pilot for Automators is highly impactful.

Process Composer Co-Pilot for Automators in Process Composer
  • Create the complete automation or edit part of the process.
  • Add a new task (Process, Bot, Form, API, Approval) to the end of the process and map dependencies.
  • Add a learning instance from Document Automation to the end of the process.
  • Set the conditions of dependencies.
  • Update supported properties of an existing node. Known limitations in Process Composer
Bot editor Co-Pilot for Automators in Bot Editor
  • Create the complete automation or edit part of the bot.
  • Add a new action to the end of the automation. When adding an action, variables for mandatory attributes are automatically created.
  • Update properties of an action.
  • Create and reference variables to edit properties.
  • Add and update the conditions of control flows (loop and if/else actions).
  • Known limitations in Bot editor

Availability

An Automator AI Platform Base license is required to enable Co-Pilot for Automators capability. A license is available to all Cloud regions. When the license is assigned, Co-Pilot for Automators capability is available to all users assigned with each of these roles and a device license:
  • Role: AAE_Basic and AAE_Bot Developer
  • Device License: Bot Creator or Citizen Developer
Note:
  • Co-Pilot for Automators is optimized for use in English. There might be limited functionality in other languages.
  • This capability is powered by an ensemble of AI/ML and large language models (LLM), and requires specific data attributes to be sent to sub-processors for completing the user requests. See Automation Anywhere Data Processing Addendum for details.
  • For more information see Data security for generative AI - FAQ.
  • For details about the Automator AI Platform Base license, see Automator AI Platform Base.

Response messages

Co-Pilot for Automators is configured to respond with standardized messages when users send prompts. These interactive messages help developers with crafting and editing prompts. Consider the following message behaviors from Co-Pilot for Automators.
  • Reject instructions not related to automation.
  • Respond with user-friendly messages.
  • Accept feedback on the output.
Automation Co-Pilot Response Message Description User Action Required
Success! Do the results meet your needs? Automation Co-Pilot has successfully created the automation, and is asking for feedback. Provide feedback by clicking Yes or No.
Automation updated. Do the results meet your needs? Automation Co-Pilot has successfully updated the process or bot, and is asking for feedback. Provide feedback by clicking Yes or No.
Thanks for your feedback. Acknowledgment for the Yes/No feedback provided by the user. None
The prompt is unrelated to Automation Anywhere capabilities. Please rephrase your prompt. The user has entered a prompt that seems unrelated to Automation 360. Correct the instruction.
Your request timed out. This could be due to high traffic. Please try again later or simplify your prompt. If the issue persists contact Support.

The error could have been caused due to time-out, API connection error or external call to the LLM model. The error is not caused due to Automation Anywhere system error. In such cases, Automation Co-Pilot retries the user request automatically before responding with this message.

 Wait for a few minutes and try again. If that does not work then reduce the message length by breaking it into smaller chunks or re-wording it.
Unable to render. Please further detail the automation scenario in your prompt. Consider using the standard names of actions and tasks. User prompt for creating or editing an automation could not be processed as Automation Co-Pilot was unable to identify the required automation entities such as tasks, packages or actions. Modify the instruction to correctly describe your automation scenario.
An unexpected error occurred, please try again. If the error persists, edit your prompt or contact Support. Automation Co-Pilot was unable to understand the request. Try rewording the instruction.
Approval task feature is unavailable. Please verify your license, role, permission and try again. Automation Co-Pilot did not have proper privileges to include this feature in the automation. Ensure the admin has granted this user the appropriate license and permissions.
API task feature is unavailable. Please verify your license, role, permission and try again. Automation Co-Pilot did not have proper privileges to include this feature in the automation. Ensure the admin has granted this user the appropriate license and permissions.
Note: Generative AI models can produce errors and/or misrepresent the information they generate. It is advisable to verify the accuracy, reliability, and completeness of the content generated by the AI model.