This topic explains the AI Skills page within AI Agent Studio, which allows you to create, update, access and use various AI Skills. This page displays AI Skills available to you from different sources.

AI Skills refer to reusable, modular components that leverages artificial intelligence (AI) models to enhance automation capabilities within the Automation 360 platform. They are pre-built AI functionalities that you can use into your automation workflows.

The Skills page is accessible from the left navigation under AI and provides a centralized view of all AI Skills accessible to you. The landing page includes two primary sections: Skills and Templates. These skills originate from the following locations:

  • Public: AI Skills that have been checked-in and made publicly available within your tenant, provided you have the necessary access permissions.
  • Private: AI Skills stored within your personal private repository.
  • Templates: AI Skill templates that have been downloaded into your tenant from the connected Bot Store. For more information, see Automation Anywhere templates.
Skills tab
This tab presents available AI Skills. You can filter the list of skills using options for name, description, and model connection name.

For each skill listed, the page displays the skill's name, the folder path where it is located, a description, the last modification timestamp, and the model connection name. You can click the folder icon on each skill to navigate directly to the skill's folder. Skills are sorted by timestamp.

You have the ability to modify private skills, including updating the associated model connection. A key feature is the option to create a new AI skill directly from this page using + Create New. For more information on creating an AI Skill, see Create AI Skills.

Templates tab
This tab lists available public templates. Templates are sorted alphabetically. You can filter the list of templates by name or description. You can create an AI Skill from a template with a bot creator license or a citizen developer license.

Creating a skill from a template involves selecting the desired template, choosing a model connection for the new skill, and saving the new skill. For more information on creating an AI Skill from a template, see Create AI Skills from a template.

Key aspects of AI Skills

  • Abstraction of AI complexity: AI Skills hide the underlying complexities of interacting with AI models. You do not need to be a data scientist or machine learning expert to use them.
  • Reusability: Once an AI Skill is created, it can be used in multiple automation workflows, saving development time and ensuring consistency.
  • Broad AI model connectivity: You have the flexibility to connect to a wide range of AI models hosted anywhere. This includes models from leading hyperscalers like Amazon Bedrock, Azure OpenAI, and Google Vertex AI, as well as open-source models (for instance, from HuggingFace or similar platforms) and even models you build and host on your own infrastructure.
  • Grounded capabilities: Many AI Skills support grounding, which means connecting the AI model to your specific knowledge bases or data stores. This helps the AI provide more accurate and contextually relevant responses.
  • Parameterization: You can configure AI Skills with parameters to tailor their behavior to specific use cases within your automation.
  • Integration with Task Bots and API Tasks: AI Skills can be seamlessly incorporated into your robotic process automation (RPA) bots and API-based integrations.

How to use AI Skills

  1. Create Model connections: First, you need to establish connections to the AI models you want to use. For instance, connect to your Azure OpenAI service.
  2. Create AI Skills: Within the Automation 360 platform, you can create AI Skills by selecting a Model connection and defining the specific task you want the AI to perform. You might configure prompts or parameters at this stage.
  3. Use in Task Bots or API Tasks: In your automation workflows, you can use the created AI Skill and provide the necessary input data. The AI Skill will then process the data using the connected AI model and return the results, which can be used in subsequent automation steps.

Get started with AI Tools

Here's an overview of the process and the role each persona plays for implementing the scenario.

Automation Admin (Jake)

Create and manage Model connections

  • Connects to a curated list of foundational models from hyperscaler vendors such as Amazon Bedrock, Google Vertex AI, Azure OpenAI, OpenAI and others.
  • Creates Standard, Fine tuned, or Grounded Model connections with RAG capability from Amazon Bedrock or Google Vertex AI.
  • Tests the model connections for viability.
  • Makes these connections available to the Pro Developers to create AI Skills via custom role.
Pro Developer (Marcus)

Create AI Skills

  • Creates an AI Skill for sentiment analysis using available Model connections.
  • Defines the AI Skill to address the business need of sorting the emails received from customers into categories and forward them to the relevant departments for further action.
  • Tests the prompt by using available Model connections, and fine-tuning the prompt parameters to optimize results for the requirement.
  • Publishes the AI Skill so the other Pro Developers and Citizen Developers can use it to create automations quickly.
Citizen Developer (Sue)

Use AI Skills in a Task Bot | Use AI Skills in API Tasks

  • Creates a new automation or updates an existing one using the AI Skill package > Execute action in the Bot editor. See: AI Skills package.
  • Chooses the AI Skill made available to her.
  • Defines the AI Skill values in the automation.
  • Runs and tests the automation to ensure all actions in the automation are running as expected.

Set up roles and permissions for all personae. See Roles and permissions.