{"id":45007,"date":"2026-06-25T16:18:26","date_gmt":"2026-06-25T10:48:26","guid":{"rendered":"https:\/\/www.inogic.com\/blog\/?p=45007"},"modified":"2026-06-25T16:18:26","modified_gmt":"2026-06-25T10:48:26","slug":"optimizing-prompt-columns-in-microsoft-dataverse-with-filters-part-2","status":"publish","type":"post","link":"https:\/\/www.inogic.com\/blog\/2026\/06\/optimizing-prompt-columns-in-microsoft-dataverse-with-filters-part-2\/","title":{"rendered":"Optimizing Prompt Columns in Microsoft Dataverse with Filters: Part 2"},"content":{"rendered":"

\"Optimizing<\/p>\n

In Part 1<\/strong><\/a>, we explored how Prompt Columns in Microsoft Dataverse bring generative AI directly into business data, helping organizations automatically summarize, classify, and extract insights from records inside their existing workflows.<\/p>\n

But as AI adoption grows, businesses need more than intelligent automation, they need control<\/strong>.<\/p>\n

Running AI on every record can quickly become inefficient, expensive, and disruptive. To address this, Microsoft has introduced an enhancement to Prompt Columns: filter-based execution<\/strong><\/p>\n

Together, this feature ensures AI only runs when specific business conditions are met.<\/p>\n

The Problem It Solves<\/strong>\u00a0<\/strong><\/h3>\n

Without execution controls, Prompt Columns can trigger AI for every record update\u2014even when the data does not require AI-generated insights.<\/p>\n

This often creates three major challenges:<\/p>\n

    \n
  1. Unnecessary AI Consumption<\/strong>
    \nAI may process blank, incomplete, or low-priority records, generating little business value.<\/li>\n
  2. Higher Copilot Credit Usage<\/strong>
    \nEvery AI execution consumes Copilot credits. Running prompts without filters can quickly increase operational costs.<\/li>\n
  3. Workflow Delays<\/strong>
    \nReal-time prompt execution across large volumes of records can slow down user actions and affect overall system performance.<\/li>\n<\/ol>\n

    With filters, organizations can eliminate these inefficiencies while maximizing AI impact.\u00a0<\/strong><\/p>\n

    Real-World Scenario: Smarter Customer Support Case Summaries<\/strong>\u00a0<\/strong><\/h3>\n

    Consider a customer support team handling thousands of service tickets every day.<\/p>\n

    The organization uses a Prompt Column to automatically generate AI-powered case summaries<\/strong> for support agents. Initially, AI was running for every ticket, including simple requests like password resets or billing questions.<\/p>\n

    This resulted in unnecessary AI usage and increased costs, while many summaries provided little practical value.<\/p>\n

    To optimize the process, the team applied a filter so the Prompt Column runs only when:<\/p>\n

    Escalation Level = High<\/strong><\/strong><\/p>\n

    This ensures AI generates summaries only for complex cases where agents need deeper context.<\/p>\n

    As a result, the organization achieved:<\/p>\n

      \n
    • Faster case resolution<\/strong> through instant access to relevant case summaries<\/li>\n
    • Reduced Copilot credit consumption<\/strong> by avoiding low-value AI executions<\/li>\n
    • Improved agent productivity<\/strong> with less manual review of case history<\/li>\n
    • Better system performance by limiting AI execution to high-value records<\/li>\n<\/ul>\n

      Note: Instead of applying AI everywhere, they focused it where it delivered the most value.<\/em><\/p>\n

      How to Implement It<\/strong><\/h3>\n

      1. Open Your Dataverse Table<\/strong><\/p>\n

      Navigate to your solution and select the table where you want AI-generated insights.<\/p>\n

      \"Optimizing<\/p>\n

      2. Create or Edit a Prompt Column<\/strong><\/p>\n

      Add a new column with Prompt<\/strong> as the data type and define the AI instruction.<\/p>\n

      Example:
      \n\u201cSummarize the customer issue and recommend next steps.\u201d<\/strong><\/p>\n

      \"Optimizing<\/p>\n

      \"Optimizing<\/p>\n

      3. Apply a Filter<\/strong><\/p>\n

      Select Apply Filter<\/strong> and define the conditions that determine when AI should run.<\/p>\n

      Examples:<\/p>\n

        \n
      • Status = Active<\/strong><\/li>\n
      • Priority = High<\/strong><\/li>\n
      • Required fields contain data<\/li>\n<\/ul>\n

        \"Optimizing<\/p>\n

        \"Optimizing<\/p>\n

        \"Optimizing<\/p>\n

        4. Monitor Execution Status<\/strong><\/p>\n

        Dataverse automatically creates Status<\/strong> and Details<\/strong> columns, allowing teams to track whether prompts are:<\/p>\n

          \n
        • Not Started<\/li>\n
        • In Progress<\/li>\n
        • Completed<\/li>\n
        • Failed<\/li>\n<\/ul>\n

          This makes monitoring and troubleshooting easier.<\/p>\n

          \"Optimizing<\/p>\n

          \"Optimizing<\/p>\n

          \"Optimizing<\/p>\n

          \"Optimizing<\/p>\n

          Final Thoughts<\/strong><\/h3>\n

          With filter-based execution<\/strong>, Prompt Columns in Microsoft Dataverse become more than just an AI feature, they become a scalable business tool.<\/p>\n

          Organizations can now control when AI runs, reduce unnecessary costs, and ensure business workflows remain fast and efficient.<\/p>\n

          The result is simple: better AI utilization, stronger ROI, and smarter business automation.<\/strong><\/p>\n

          Frequently Asked Questions (FAQs)
          \n<\/strong><\/h3>\n

          1. What are Prompt Columns in Microsoft Dataverse?<\/strong><\/p>\n

          Prompt Columns in Microsoft Dataverse are AI-powered columns that use generative AI to automatically summarize, classify, or extract insights from data stored in tables. They help embed Copilot-like intelligence directly into business records and workflows.<\/p>\n

          2. How do filter-based Prompt Columns work in Dataverse?<\/strong><\/p>\n

          Filter-based Prompt Columns allow you to control when AI is triggered. Instead of running on every record, the prompt executes only when defined conditions (such as Priority = High or Status = Active) are met, improving efficiency and reducing unnecessary AI usage.<\/p>\n

          3. Why should I use filters with Prompt Columns in Microsoft Dataverse?<\/strong><\/p>\n

          Using filters helps organizations:<\/p>\n

            \n
          • Reduce Copilot credit consumption<\/li>\n
          • Avoid running AI on irrelevant or incomplete data<\/li>\n
          • Improve system performance<\/li>\n
          • Ensure AI is used only where it adds business value<\/li>\n<\/ul>\n

            4. Do Prompt Columns in Dataverse consume Copilot credits?<\/strong><\/p>\n

            Yes. Every time a Prompt Column executes, it consumes Copilot or AI credits depending on your Microsoft licensing model. Applying filters helps significantly reduce unnecessary credit usage.
            \n