Today, organizations rarely struggle with collecting data. In fact, most companies already have access to more data than ever before from ERP systems, CRM platforms, web applications, local files, cloud environments, and enterprise data platforms such as OneLake. The real challenge begins after the data has been collected: how to connect it, transform it, standardize it, and turn it into insights that support better business decisions.
This was one of the key themes discussed during FabCon 2026 in Atlanta, where Marcin Kolman presented the role of DataFlow Gen2 in Microsoft Fabric. His session focused on a practical question: how can organizations simplify data transformation and accelerate the path from raw data to meaningful analysis?
Data is available — but is it truly usable?
Modern organizations generate and store data across many systems, but availability alone does not create business value. Data often comes in different formats, structures, and levels of quality. It may require manual preparation before it can be analyzed, and transformation processes are often repeated again and again — for example, as part of monthly reporting cycles.
In practice, this means that teams spend valuable time preparing data instead of analyzing it. Instead of focusing on trends, risks, opportunities, and business outcomes, they are forced to repeat technical or semi-manual tasks that slow down decision-making.
An ideal data preparation process would look very different. Connecting to data would be simple and intuitive, even for less technical users. Data would be ready for analysis faster, transformations would be reusable rather than one-off, and refresh processes would not require constant manual intervention.
This is exactly the direction in which Microsoft Fabric is evolving — and DataFlow Gen2 plays an important role in that vision.
DataFlow Gen2 as a modern ETL foundation
DataFlow Gen2 supports the full ETL process: extracting data from different sources, transforming it, and loading it into a selected destination. What makes it especially accessible is the fact that it is built on Power Query — a tool already familiar to many Excel and Power BI users.
This matters because it lowers the barrier to entry. Business users and analysts who have already worked with Power Query in Excel or Power BI can move into the Fabric environment without completely changing the way they work. At the same time, DataFlow Gen2 is not limited to local desktop processing. Transformations are executed using Fabric resources, which makes them more scalable and better suited to larger data volumes. [
As a result, DataFlow Gen2 becomes more than a simple preparation tool. It can become a reusable part of a broader data management process, where transformations are created once and then used consistently across reports, models, and analytical workflows.
Easier integration across many data sources
One of the most common challenges in data work is integration. Data often lives in many systems that do not naturally communicate with one another. DataFlow Gen2 helps address this challenge by offering more than 170 built-in connectors, broad support for custom connectors, and integration with OneLake and other Fabric components.
This means users can connect to multiple sources from one environment and prepare data for analysis without constantly switching between tools or relying on separate manual workflows. For organizations working with diverse data ecosystems, this kind of integration is essential for building a more consistent analytical foundation.
From one-time transformations to reusable processes
In many companies, data transformation still happens as a repetitive, manual process. Data is exported, cleaned, adjusted, and then the same steps are repeated during the next reporting cycle. This creates unnecessary workload and increases the risk of inconsistency.[
DataFlow Gen2 changes this approach by allowing transformations to become repeatable and centrally managed. Instead of preparing data separately for each report, teams can create transformation logic once and reuse it in different analytical scenarios.
This is particularly important when working with Power BI. A new capability allows users to export query results and transformations from Power BI Desktop directly into DataFlow. This helps move transformations out of a single report and into a more standardized data process that can support multiple reports or models. [
New destinations: SharePoint Excel files and Snowflake
DataFlow Gen2 is also expanding the range of destinations where transformed data can be loaded. Recent updates include support for saving query results to Excel files in SharePoint and Snowflake databases.
This creates more flexibility for teams that need to deliver data in formats and environments already used across the organization. For example, query results can be exported not only as a table, but also directly into a chart in an Excel file, helping users move from transformation to analysis more quickly.
For Snowflake, the process is designed to be straightforward: after creating a connection, users select the connection, choose the schema, name the target table, and load the transformed data directly into the database.
Faster preparation with “table from examples”
One of the most interesting capabilities discussed in the session was the option to create a table from examples. Instead of manually building a complex set of transformations, users can provide a few sample values, and DataFlow Gen2 fills in the rest based on the pattern.
This can be especially useful when working with less structured data, such as text files exported from external systems. In these cases, users often need to transform messy information into a clean tabular format. With examples, the system can generate the necessary Power Query logic in the background, reducing the number of manual steps required. ]
The value of this feature is not only technical. It also makes data preparation more accessible to people who may not know exactly how to write or configure every transformation step manually.
AI-powered transformations with Copilot and prompts
DataFlow Gen2 is also becoming more intelligent through AI capabilities. Copilot is available in DataFlow Gen2 and can help users generate data or create transformations using natural language.
A practical example discussed in the session involved customer feedback. When a dataset contains a column with customer opinions about a product or service, AI can help classify each comment as positive, neutral, or negative. Users can describe the desired outcome in natural language and provide the relevant column as context.
This kind of functionality can significantly accelerate analytical work. Instead of manually designing every transformation or classification rule, users can explain what they want to achieve and let AI support the creation of the required logic.
Variable library: more consistent logic across processes
Another important feature is the ability to use values from a variable library. This allows organizations to store global variables and manage key parameters centrally.
For example, if several processes rely on the same filtering threshold, that value can be maintained in one place instead of being changed separately across multiple components. This improves consistency and makes it easier to update business logic when requirements change.
For larger organizations, this is especially valuable because many teams may work with the same data, rules, and assumptions. Centralized logic helps reduce fragmentation and supports better governance of analytical processes.
What does this mean for business?
DataFlow Gen2 is not just a technical feature inside Microsoft Fabric. It represents a shift in how organizations can approach data preparation and transformation. By simplifying integration, automating repetitive work, supporting reusable transformations, and introducing AI-assisted capabilities, it helps teams move faster from raw data to business-ready insights.
For business users, this means less time spent on manual preparation and more time available for analysis. For data teams, it means more scalable, standardized, and repeatable processes. And for the organization as a whole, it means a shorter path from data collection to informed decision-making.
In a world where companies are collecting more data every day, the competitive advantage no longer comes from simply having access to information. It comes from being able to transform that information quickly, consistently, and intelligently. DataFlow Gen2 in Microsoft Fabric brings organizations closer to that goal
