Skip to content
  • There are no suggestions because the search field is empty.

AI Studio example prompt – Generating stock item descriptions

Learn how to create an AI action that automatically generates stock item descriptions.

When creating a new stock item, your users might take a long time to write an appropriate description. Because they are already entering attributes for the item, you can create a prompt that automatically generates the description based on those attributes.

  1. Go to the LLM Prompts form (ML202000).
  2. In the summary area, enter some basic details about the prompt you're creating:
    • Prompt ID  and Prompt Name – Enter a descriptive name for the prompt. This is how you'll find it again later on the LLM Prompts form. In this example, let's make the ID STOCK-ITEM-DESC and the name Stock Description from Attributes.
    • LLM Connection – Which large-language model the prompt uses. The default connection is a model managed by MYOB. You can also set up more connections on the LLM Connections form (ML201000).
    • Source Form – The form where the action button will be added. In this example, we'll add an action to the Stock Items form (IN.20.25.00).
    • Button Name – How the action button is labelled on the source form. In this example, let's label the button Generate Description.MACU_AIStudio_Example1Summary
  3. On the Instructions tab, describe the specifics of what you want the AI agent to do when a user clicks the prompt button. This tab is made up of three sections: ## Context Instructions, ## Instructions with Input Data and ## Output Data Field. Make sure you leave these section headings in when you add your instructions. The table below provides example instructions you can copy and paste.
    Instructions tab section Example instructions
    In the ## Context Instructions section, tell the AI agent how to act.

    ## Context Instructions

    You are an AI assistant working with Acumatica ERP data, functioning as an experienced marketing copywriter specialising in product descriptions.
    In the ## Instructions with Input Data section, describe what you want the AI agent to do with your inventory data.

    ## Instructions with Input Data

    Transform raw inventory data from item attributes into a polished product description:
    Short Description: Concise, corrected item description.

    Inventory Attributes:

    <foreach view="Answers">

    • ((Answers.AttributeID)):((Answers.Value))

    </foreach>
    In the ## Output Data Field, choose  which fields you want the AI to add its stock descriptions to.

    ## Output Data Field

    Short Description: @((Item.Descr))

    Format: Plain text string

    Length: Maximum 50 characters

    Purpose: Quick item identification for internal systems and brief listings

    Requirements:

    • Fix grammar, spelling, and punctuation errors
    • Ensure clarity and specificity
    AIStudio_GenerateStockDescriptionInstructions
  4. On the Output Format tab, enter an example of what the AI agent should write when a user clicks the prompt button. Because the example needs to be formatted in a specific way, we recommend clicking Generate Example and then fine-tuning it.
    For this prompt, you can copy and paste the following output format:
    {
    "((Item.Descr))": "Intel Core i7: 8 cores, 3.6–5.0 GHz clock speed, 16 MB cache"
    } AIStudio_GenerateStockDescriptionOutputFormat
  5. After entering your instructions and output format, save your changes.
  6. In the Prompt Testing section, select an example record to run the prompt on and click Test Prompt.
  7. If you're happy with the results in the Preview tab, select the Active checkbox in the summary area. This adds the prompt button to the source form you selected in step 2. Your users can now take advantage of the prompt to streamline their workflow.