AI Tools
- Última actualización2024/10/03
AI Tools
Enhance your solution with AI Agent Studio. This capability addresses a variety of business requirements where a predefined AI Skill can be used to streamline an action and receive a relevant response in an automation, helping reduce manual effort substantially.
Overview
Use this feature to build generative AI-infused intelligent automations that span your organization. Ensure use of approved foundational models within your solutions by creating Model connections that can be shared with specific users, teams or departments. Evaluate different models and fine-tune prompts that are optimized for your use case. The combination of a foundational model and a fine-tuned prompt is called an AI Skill, which can be made available to users to accelerate automation development and achieve a higher level of productivity by building automations quickly and easily.
The following video gives you an overview of AI Tools with a focus on its value offering:
Benefits
- Connect to a curated list of foundational models from hyperscaler vendors such as Azure OpenAI, OpenAI, Amazon Bedrock, and Google Vertex AI.
- Make these connections available to the Pro Developers to create AI Skills.
- Secure user instances to ensure processing within user-governed tenant, with complete data encryption.
- Create AI Skills using available Model connections.
- Define AI Skills as per your requirements.
- Test prompts using different models (available to you) and fine-tune the prompt parameters to optimize results for your use cases.
- Publish AI Skills for use across automations within your organization.
- Use AI Skills made available to you in the AI Skills package in the Bot editor.
- Create or update your automations with AI Skills.
- Define or configure AI Skill inputs in your automation.
- Run your automation to test and ensure all actions in the automation are running as expected.
- Drive stronger governance with an ability to monitor and audit AI usage within automations.
- Ensure use of approved models by developers.
- Enforce compliance and regulatory policies.
The following visual gives you an overview of the AI Tools architecture. Refer to the table below the diagram for details on the call-outs.
Model connections and AI Skills are designed and stored within the Control Room. | |
When a prompt is executed, it gets forwarded to a local execution agent within the customer infrastructure. | |
A prompt is sent to the customer’s chosen foundational model and returned directly to the customer infrastructure. | |
All prompts, responses, model connection details, and any other additional information is stored within the AI governance logs in the customer environment. |
Availability
An Enterprise Platform license is required to use AI Tools capability for AI Agent Studio.
For information about supported version for this feature, see Enterprise Platform