Data Warehouse: Copilot & AI Skill
AI is transforming data warehousing, making complex analytics more accessible and efficient. With tools like Copilot for Data Warehouse and AI Skill, Microsoft Fabric offers two powerful, complementary resources that serve both data developers and business users. This blog explores how these tools differ, when to use each, and how they can work together to maximize productivity and deliver insights with Fabric Warehouse. Let’s start with a brief introduction to each feature.
What is Copilot for Data Warehouse and AI Skill
Copilot for Data Warehouse
Copilot acts as a helpful assistant, guiding users through the complexities of data warehousing. For those familiar with ChatGPT, think of Copilot as a specialized AI that’s ready to help you build warehouse schemas, generate SQL code, and extract key insights from data— all directly integrated in the warehouse editor. Designed for both technical users and business professionals, Copilot acts as your own on-demand programmer & data analyst, ready to simplify and speed up your data discovery process.
AI Skill
AI skill is a tool that allows you to configure a generative AI assistant that provides answers and insights into your warehouse data by generating queries. With the ability to configure an AI skill, users can add additional business context and guide the AI on how to respond to questions. Once configured, it allows users to make data accessible by generating queries in response to plain-language questions. Notably, AI skill is especially useful for creating customized, shareable interface to talk to warehouse data – inside or outside of Fabric – that can adapt to specific organizational needs.
When Should I Use It?
Although Copilot and AI skill both generate SQL queries, the table below summarizes some key aspects to consider when choosing between the two.
| Feature | Copilot for Data Warehouse | AI Skill |
| Purpose | Paired programmer; Provides interactive, real-time assistance across tasks within Microsoft Fabric Warehouse (not just SQL queries)! | Domain expert; Generates context-tailored SQL queries in response to natural language questions for customized data access. |
| Configuration | General use; user-configurable customizations through prompts. | Highly customizable; can be configured with instructions and SQL examples to fit specific use cases. |
| Prompting | Interact with natural language or directly through SQL code. | Interact with natural language. |
| Scope | Assists with a wide range of Fabric tasks, including code generation, documentation, troubleshooting, and general warehousing help. All types of queries are supported: DML, DDL, and DQL. | Focuses on read-only queries for structured data. Only DQL queries are supported. |
| Integration | Limited to in-app Fabric assistance; no cross-tool capabilities. | Flexible; Use inside or outside of Fabric via published URL. |
| Use-Case | Supports immediate help with query composition, code snippets, warehouse Q&A, and data exploration. | Ideal for creating self-serve query/Q&A tools tailored to organizational datasets. |
| Best Suited For | Real-time support for warehouse development, analysis, and guidance on best practices. | Natural language Q&A on top of warehouse data. |
As you can see, Copilot and AI skills are complementary features when developing and analyzing a data warehouse. Copilot is best used as a real-time, interactive, assistant when building and exploring the warehouse. After foundational data prep and warehouse development is complete, AI skill can be created and configured to make warehouse data accessible and available for everybody in your organization – regardless of pro-code or no-code scenarios. Using both Copilot and AI skills together in your warehousing solutions is a recipe for success and is highly recommended.
Better Together
Now that we have a high-level understanding of the two features and use-cases, let’s dive deeper into how the two play complementary roles in a developer’s workflow when building and interacting with data warehouses.
For data warehouse developers and business users alike, combining the strengths of AI skill and Copilot can lead to a more powerful, streamlined experience:
- Data Exploration: AI Skill can generate tailored SQL queries that reveal patterns or critical data points, which Copilot can then help refine or modify within a real-time context. For instance, a Data Analyst might use an AI Skill to pull complex metrics from a dataset, and then turn to Copilot to iteratively adjust the parameters or conditions based on initial insights.
- Enhanced Collaboration: By configuring AI skill for broader use across teams, developers can provide non-technical colleagues with tools to query data without SQL knowledge. Meanwhile, Copilot supports developers directly in their workflows, assisting with debugging, optimizing, or expanding queries generated by AI Skill.
- Reduction of Errors and Improved Efficiency: AI skill is highly configurable to produce contextually accurate outputs for specific datasets, which reduces the risk of misinterpreting user queries. When paired with Copilot’s guidance, it’s easier to troubleshoot or verify these results, ensuring both tools contribute to more reliable decision-making.
For example, a Developer working on a data warehouse project can use Copilot to quickly set up the database schema and write initial queries. Then, AI skill can be employed to analyze the data and provide insights that inform further development.
By leveraging AI skill for creating accessible, self-serve queries, and Copilot for task-specific support, data warehouse users can streamline both simple and complex workflows to make data accessible for all in Microsoft Fabric.
Next Steps
Interested in learning more?
Join our livestream, November 7th, to learn more about how to use AI skill and Copilot effectively together in the Fabric Warehouse.
Get started today
Get started with Copilot for Data Warehouse: Copilot for Data Warehouse (preview) – Microsoft Fabric | Microsoft Learn
Get started with AI skill in Microsoft Fabric: AI skill creation (preview) – Learn how to create an AI skill | Microsoft Learn
Have a feature request? Post your thoughts to Ideas (microsoft.com).
Join us in this exciting journey to make data more accessible, understandable, and actionable for everyone in your organization. Thanks for reading!
