Hello friends! 👋
If you're working with Microsoft Power Platform, especially Dataverse, here’s a cool update you should know about – a new column type called "Prompt" is now available (in preview as of writing this post).
This is one of those features that make you go, “wow!
Finally, this is here.” So in this article, I’ll walk you through:
- ✅
What is a Prompt column in Dataverse?
- 🎯
What’s the purpose or use case?
- ⚙️
How to create a Prompt column?
- 🧑💻
One real-time example (simple and relatable)
Let’s dive in!
🚀 What is a Prompt Column
in Dataverse?
The Prompt column type is a new data type
introduced in Microsoft Dataverse (preview) that lets you trigger
AI-based prompts from within model-driven apps.
In simpler terms: it’s a column that holds a prompt template
for Copilot-style AI responses.
📌 This new column allows Copilot
(AI) to generate answers or outputs based on row data, using the
prompt configured inside the column.
🤔 Why Use Prompt Columns?
(Purpose)
The main purpose of Prompt columns is to bring
generative AI capabilities closer to your business data.
Here are a few practical use cases:
- 📄
Auto-generate a product summary from product details
- 📧
Create customer response emails using ticket data
- 📊
Generate insights from case records
- 🧾
Draft invoice remarks using line items
Instead of building complex Power Automate flows or calling
external APIs, you can just store the prompt directly in a column and
use it via Copilot.
That’s right – AI becomes a native part of your Dataverse
schema!
🛠️ How to Create a Prompt
Column in Dataverse?
Let me walk you through the steps (assuming you're familiar
with Power Apps Maker Portal):
🔹 Step 1: Go to Power
Apps > Tables
Open Power Apps Maker Portal → Choose your environment →
Click on Tables.
🔹 Step 2: Choose a Table
or Create a New One
For example, you can pick a Case table or create your
own “Customer Feedback” table.
🔹 Step 3: Add a New
Column
Click “+ New Column” → Under Data Type, select
Prompt (Preview).
🧠 Note: If you don’t see
Prompt column type, make sure AI Builder is enabled and you're in a
supported region.
🔹 Step 4: Configure
Prompt Template
Inside the column configuration, you’ll define a Add prompt.
This template can include placeholders from other columns of the row.
For example:
Summarize the customer issue based on the case title: and
description:
This template will be used by Copilot to generate responses
dynamically.
🔹 Step 5: Save &
Publish
Save your column and publish the table.
That’s it – your Prompt column is ready!
✅ Example: Generate Response
Summary for a Case Record
Let’s say you have a Case table with the following
columns:
- Title:
"Unable to login"
- Description:
"User reports login error on mobile app."
You create a Prompt column with this template:
When Copilot reads this, it will return something like:
" The customer issue involves an employee
named Diana Prince who is experiencing a login error on the mobile app. The
case is currently active, and the user has reported being unable to log in. The
case is owned by Gowtham Rajamanickam and was created on May 21, 2025. The
status of the case is active, indicating that the issue is still being
addressed.
Boom! ✨ You’ve
automated a contextual AI response without writing any code.
🧠 Things to Remember
- 🧪
Prompt columns are currently in preview
- 📍
You can refer to other columns using curly braces {columnname}
- 🧱
Works only in Model-Driven Apps
- 🧾
It’s read-only, used for AI input – not user-entered values
📚 Microsoft Docs
Reference
If you want the full official reference, check this out:
🔗
Prompt columns (preview) – Microsoft Docs
👨💻 Final
Thoughts
The Prompt column in Dataverse is a real game-changer
🔥
for developers and makers working with Power Platform and Copilot.
It bridges AI with structured data beautifully. And
it fits right into the low-code/no-code vision Microsoft has been pushing for
years.
Whether you're a Power Platform Pro or just getting started,
I highly recommend exploring Prompt columns — especially if your apps could
benefit from context-aware, dynamic AI generation.
Let me know in the comments if you've tried it already or
have ideas to implement it in your solutions.
No comments:
Post a Comment