Copilot Overview
In the Data Productivity Cloud, Copilot introduces a new era of AI-driven data productivity tools, letting you prompt generative AI to create data pipelines, improving your efficiency in managing and processing data.
With Copilot, you can create data pipelines simply using plain language instructions, taking advantage of advanced AI technology. And Copilot accompanies you throughout your data pipeline journey, offering help at every step.
Note
- Currently, Copilot is only available for Snowflake projects.
- Currently, Copilot is only available for transformation pipelines.
Note
Copilot is still in its early stages and is considered experimental. While it offers promising features, it's continuously evolving as we gather feedback and refine its capabilities.
Important considerations
- Exclusivity: Copilot is currently available exclusively for Transformation pipelines.
- Ongoing refinement: We are continuously enhancing Copilot's capabilities, especially in areas such as configuring table inputs, managing multi-table inputs, handling flattened variants, and extracting nested data columns that contain JSON data.
- Potential challenges: Users may sometimes find that Copilot places components incorrectly on the canvas or suggests irrelevant inputs. We are dedicated to resolving these issues and improving component configurations to enhance clarity and reduce errors in the automation process.
Note
While Copilot can add and update components on the canvas, it's unable to perform other actions such as delete component, undo action, or run and schedule pipeline.
Although Copilot represents a promising advancement in AI-driven data productivity, it's important to understand that it's still in its experimental phase. We encourage you to use Copilot's feedback feature as we continue to improve and enhance its capabilities.
The flow of sending data to Copilot
Your interaction with Copilot involves transmitting instructions for processing. This flow encompasses the following steps:
- Users provide instructions and prompts directly within the Copilot chat interface to guide the data pipeline creation process.
- Copilot uses user's input along with its knowledge of Matillion's Data Productivity Cloud and common data transformations to fulfill user requests. For example, it can select components for filtering, aggregation, and other operations needed to create the right data pipeline. However, Copilot does not access the data stored in your data warehouse; it is only aware of the pipeline configuration present on the canvas, including the column metadata.
- Copilot generates metadata for transformation components, including details such as column names, data types, sizes, and precision levels. This metadata is derived from the user's instructions.
Before you start using Copilot, read Copilot prerequisites.