Create AI Skills with Grounded by data store Model connections

This is the next logical step after creating a Grounded by data store Model connection. You would create an AI Skill and connect it to a Grounded Model connection from Google Vertex AI.

주:

Google Vertex AI Grounded by data store Model connections is now available in Automation 360 v34 release on Cloud as well. You can use this feature in Cloud and On-Premises.

Google Data Store with document chunking is now supported to ensure optimal results in automation executions. You can enable document chunking in the Google Data Store for using Google Vertex AI Grounded models in AI Agent Studio.

A Pro Developer creates AI Skills so the Bot Creators can use these in their automations to save time and effort.

AI Skills are created by connecting to Model connections the Pro Developer has access to. Next, the AI Skill is fine-tuned by testing the prompts using different grounded models to find the best response that addresses the business ask. These AI Skills are made available to Bot Creators for use to help accelerate the task of creating automations across solutions.

전제 조건

A Pro Developer requires these roles and permissions to create and test AI Skills.
  • Role: AAE_Basic, Pro Developer Custom role
  • Permission: Bot Creator

See 역할 및 권한

Other requirements:

Besides the roles and permissions the Pro Developer must be connected to a Bot Agent 22.60.10 and later. As part of testing the Model connection, you would have to run the bot on your desktop. Hence ensure the Bot Agent is configured to your user. If you have to switch connection to a different Control Room, see: Control Room 인스턴스 간 기기 등록 전환.

프로시저

  1. Log in to the Control Room and navigate to Automation > Create new or ‘+’ icon and choose AI Skills.
  2. Provide a name and description and click Create & edit to display a template outline.
  3. In the AI Skills screen, click Choose model connection to choose from the available list of Model connections you have access to. You would choose a Model connection that was created using the Grounded by data source type option from Google Vertex AI.
    These Model connections are created by the Automation Admin and assigned to your user with a custom role.
  4. After selecting a Model connection, the AI Skills is set up with the default parameter settings that is optimal for the chosen model. You can change the settings as required.
    The AI Skill editor displays with default parameter values set by the model vendor which you can configure as required. These values can be configured when creating a data store in Agent Builder.
  5. Next, add a filter condition, which is optional. This field supports a string format for entering the filter value. Adding a filter helps narrow down the model's search to the specific files within the Google Data Store.

    You can make sure that the AI Skill grounds the response using information in a specific set of documents in the Google Data Store by adding relevant metadata to the data store. This narrows down the scope of response and makes it more accurate. This also helps reduce the size of the data retrieved from the vector data store before passing it to the foundational model. If the data store content is vast, this filtering mechanism using metadata helps get a better response performance from the Prompt query. For example: for the example What is the gift tax limit for the year 2024? (see below), you could use a filter for tax_year: ANY("2024").

    See Filter generic search for structured or unstructured data .

  6. Now you can start creating an AI Skill and add prompt inputs, as required. Let us use an example to walk you through the steps.
  7. In the Prompt field enter your Prompt text with the input variables.

    For example:

    What is the gift tax limit for the year 2024?

    Prior to this step, you would have uploaded tax rule PDF documents for the last 3 years in the Agent Builder such as: tax_rules_2022.pdf, tax_rules_2023.pdf, and tax_rules_2024.pdf.

    The response for the Prompt text will be referenced from the tax_rules_2024.pdf document.

  8. Click out of the Prompt entry field.
  9. Click Get response to get a response from the model based on your Prompt.
    주: 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.
  10. The Grounded by data store Model connection returns a response in the Response field, and additionally displays a Citations field displaying all citation references.

    Citations are chunks of information stating, which section of a document stored in the Grounded by data store, the response is referenced from. You can see the document title of the referenced data store from the Google Data Store.

    See Parse and chunk documents , and Chunk documents for RAG .

    주: The Document retrieval count parameter determines the maximum number of chunks that can be retrieved from the data store and provided as context to the hyperscaler model to generate the response.

다음 단계

Your next step would be to check-in the AI Skill to make it available to Citizen Developers and Pro Developers using the AI Skills package.

Why would you check-in an AI Skill?

After creating an AI Skill, check it into the Public folder. This will let the Pro Developer, and Citizen Developer use it from the AI Skills package in the production environment.

For reference, see:

A Task Bot, with one or multiple embedded AI Skills can be added to a larger automation that would run a complete workflow scenario. You would create such a workflow in a Process Composer.

주: When you create or test an AI Skill in the AI Skill screen, the success or failure details along with the model responses can be viewed in these navigation screens:
  • Administration > AI governance > AI Prompt log
  • Administration > AI governance > Event log
  • Administration > Audit log

See AI Governance.

As the next step in your sequence of tasks, go to AI Skills에 Task Bot 사용 and use the AI Skill in an automation.