TL;DR
Data Factory in Microsoft Fabric is the built-in integration layer that connects to 170+ data sources, moves and transforms data at scale, and orchestrates entire pipelines from inside Fabric’s unified platform, removing the need for a separate integration tool alongside your analytics stack; existing Azure Data Factory users get a largely familiar experience, with the main differences showing up in how pipelines connect to OneLake and Fabric’s other workloads.
Picture a mid-size company where finance data lives in SQL Server, customer records sit in Salesforce, and operational data is spread across an on-premises database that nobody wants to migrate. The analytics team spends more time wrangling data than actually analyzing it.
This is not an edge case. It is the default state for most organizations that have grown organically, acquired tools over time, or inherited systems from past IT decisions. And the cost of that fragmentation shows up everywhere, in slow reporting cycles, unreliable dashboards, and decisions made on stale numbers.
Data Factory in Microsoft Fabric is the integration layer built to change that. It connects to 170+ data sources, moves data at scale, transforms it, and orchestrates the entire workflow from one place inside Fabric’s unified platform .
In this article, we’ll cover what Data Factory in Microsoft Fabric actually is, its key features and capabilities, real-world use cases, and what existing Azure Data Factory users need to know before making the switch.
Key Takeaways Data Factory is Microsoft Fabric’s built-in service for connecting, moving, and transforming data at scale It connects to 170+ sources across on-premises, cloud, and multi-cloud environments Supports both ETL and ELT , and teams can combine both in the same pipeline Dataflow Gen2 lets analysts build transformation workflows without writing a single line of code Copilot is built in, so teams can design pipelines and fix errors using plain-language prompts
What Is Data Factory in Microsoft Fabric? Data Factory is one of six core workloads inside Microsoft Fabric , the platform Microsoft built to bring data engineering, analytics, real-time intelligence, and business intelligence together in one environment.
All of these workloads sit on top of OneLake, Fabric’s centralized data lake . That matters for Data Factory specifically because pipelines you build do not output data to a separate storage system. The results land directly in OneLake, where every other Fabric workload can access them without copying data or managing connectors between services.
Its core job is straightforward: connect to data sources, move data, transform it, and orchestrate workflows. The difference from standalone integration tools is that everything happens inside the same platform your analytics and reporting teams already use.
A note for Azure Data Factory users: Fabric Data Factory is the next generation of ADF, not a competing product. Existing ADF pipelines, connectors, and patterns carry forward, with meaningful additions on top.
So that is what it is. Now for how it actually functions under the hood.
How Data Factory in Microsoft Fabric Actually Works Data Factory is built around four core capabilities. Understanding how they fit together makes the rest of the feature set easier to evaluate.
Before getting into each one, it also helps to know that Fabric supports both ETL and ELT patterns. ETL transforms data before it reaches the destination, which works well when clean, structured data is needed on arrival. ELT loads raw data first and transforms it in place using Fabric’s compute engines, which suits larger datasets where cloud-scale processing handles the heavy lifting. Teams can combine both in the same solution.
1. Connect Data Factory pulls from 170+ native sources, spanning on-premises systems, cloud platforms, SaaS applications, and multi-cloud environments. Hybrid setups are supported through on-premises data gateways and virtual network gateways, so legacy systems do not need to be migrated before they can be connected.
Azure SQL, SharePoint, Salesforce, SAP, Snowflake, and 165+ other sources supported natively On-premises connectivity through data gateway or virtual network gateway Multi-cloud support covers AWS, Google Cloud, and hybrid environments in the same pipeline 2. Move Getting data from source to destination is handled through three purpose-built options, each suited to different scenarios. Choosing the right one depends on your volume, latency requirements, and how much control the workflow needs.
Copy Job Copy Activity gives teams fine-grained control over parallelism and transformation logic during the copy process Mirroring creates near real-time database replicas inside OneLake, supporting SQL Server, PostgreSQL, Cosmos DB, and Snowflake Raw data rarely arrives in a format analytics tools can use directly. Data Factory handles transformation through both low-code and code-first options, so analysts and engineers can both contribute without getting in each other’s way.
Dataflow Gen2 offers 300+ visual transformations including joins, aggregations, and data cleansing , no code required dbt job brings SQL-based transformations natively into Fabric for teams already running dbt workflows Pipeline activities support notebooks, Spark job definitions, stored procedures, and SQL scripts for custom transformation logic 4. Orchestrate Pipelines in Data Factory go beyond simple linear sequencing. They handle real workflow logic, including loops, conditionals, branching, and error handling, so complex multi-step processes can be expressed as a single managed workflow.
Schedule-based and event-driven triggers supported, including file arrival, folder events, and pipeline completion events Variable libraries (in preview) allow teams to reuse the same variables across multiple pipelines in a workspace 5. Copilot Copilot is embedded directly in Data Factory and supports pipeline design, dataflow authoring, and error diagnosis using plain-language prompts. Users describe what they need, and Copilot generates the corresponding expressions or transformation steps.
Summarizes existing dataflow queries and pipelines so teams can understand inherited logic quickly Diagnoses failed pipeline runs with categorized issues, root cause analysis, and suggested fixes
The pipeline expression builder Copilot is currently in preview. The broader Copilot integration across pipelines and dataflows is generally available.
That flexibility across all five capabilities is what the features section below builds on.
Key Features of Microsoft Fabric Data FactoryThe five capabilities above describe what Data Factory does structurally. The features below are what make each of those capabilities useful at enterprise scale.
1. 170+ Native Connectors, Including On-Premises and Multi-Cloud Data Factory connects to databases, file systems, APIs, cloud platforms, SaaS applications, and on-premises systems without requiring a separate integration layer. For teams managing hybrid environments, this means both on-prem and cloud sources can feed the same pipeline.
Covers Azure SQL, SharePoint, Salesforce, SAP, Snowflake, and 165+ other sources On-premises connectivity through data gateway or virtual network gateway Multi-cloud support for AWS and Google Cloud alongside Azure sources 2. Three Data Movement Options for Different Scenarios Copy Job handles most standard scenarios with bulk, incremental, and CDC delivery modes, and now auto-creates destination tables on the fly. Copy Activity gives engineers fine-grained control over parallelism and copy behavior. Mirroring supports near real-time database replication into OneLake without a traditional ETL pipeline .
Copy Job: simplified movement with bulk, incremental, and CDC support, no manual schema setup needed Copy Activity: manual control over parallel copying, transformation logic, and custom configurations Mirroring: near real-time replication from SQL Server, PostgreSQL, Cosmos DB, and Snowflake into OneLake Dataflow Gen2 gives analysts a visual interface to clean, reshape, and transform data without writing code. It handles joins, aggregations, data cleansing, and custom logic through 300+ built-in operations, and supports Python and R for teams that need more flexibility.
300+ built-in transformations accessible through a drag-and-drop interface Preview-only steps let teams validate logic against sample data without affecting production Teams already running dbt workflows can now run them natively inside Fabric Data Factory without rebuilding from scratch. dbt models are authored, orchestrated, and deployed inside Fabric, with access to Fabric’s governance and monitoring on top.
dbt jobs run natively inside Fabric as of public preview announced at Ignite 2025 Combines dbt’s version-controlled, testable approach with Fabric-native governance and CI/CD Supports analytics engineering teams that prefer SQL-based modular transformation workflows 5. Pipeline Orchestration with Full Control Flow Logic Pipelines support loops, conditionals, branching, and error handling, covering complex multi-step workflows in a single orchestrated process. Schedule-based and event-driven triggers let pipelines start automatically based on time, file arrivals, or the completion of another pipeline.
Supports loops, conditionals, branching, and error handling for complex workflow logic Event-based triggers on file arrival, Lakehouse folder events, and pipeline completion Apache Airflow integration available for teams building DAG-based orchestration in Python Microsoft Fabric Data Agents: How They Work in 2026 Learn how Microsoft Fabric Data Agents turn natural language questions into real-time data answers
Learn More
What This Means for the Businesses? Key Benefits Features matter. What organizations care about in practice is what those features change day to day for the people running the data function. Here is where the impact shows up.
When data integration , transformation, storage, and analytics run inside one platform, teams stop spending time maintaining connections between separate systems. The overhead of keeping ADF, Synapse, and standalone visualization tools in sync largely goes away.
Replaces separate integration, compute, and analytics tools with one Fabric capacity subscription Reduces pipeline maintenance caused by schema changes or API updates in disconnected systems Simpler onboarding for new data team members who only need to learn one environment 2. Analysts and Engineers Can Both Contribute Dataflow Gen2’s low-code interface lets analysts build and maintain transformation logic without waiting for an engineering ticket. Engineers retain full access to notebooks, Spark, SQL scripts, and dbt when more complex logic is needed.
Analysts manage their own data prep without creating a backlog in the engineering queue Both roles work inside the same platform, reducing handoff friction between teams 3. Governance Comes Included Every pipeline, transformation, and dataset in Fabric inherits the same governance model automatically. Sensitivity labels, Purview integration, lineage tracking, and role-based access are applied at the platform level rather than configured per tool.
Sensitivity labels and data classifications persist across all connected workloads Lineage tracking covers the full journey from source system to Power BI report Regulated industries get a compliance-ready governance structure without a separate setup 4. Cost Structure Simplifies Instead of licensing separate integration, compute, and visualization services, organizations pay for Fabric capacity that covers all workloads. For teams currently running ADF, Synapse, and Power BI Premium separately, consolidation onto Fabric changes both the cost model and the management overhead.
Eliminates overlapping licensing between ADF, Synapse, and standalone ETL tools Operational spend on infrastructure management decreases as workloads consolidate onto one platform
Where Teams Are Actually Using Fabric Data Factory? Top Use Cases Knowing what a tool can do is one thing. Seeing where it fits into real workflows makes it easier to evaluate whether it is the right fit for your situation. Here are five scenarios where organizations are using Data Factory in Microsoft Fabric today.
1. Centralizing Multi-Source Data for Reporting A finance team needs a consolidated weekly view pulling from an ERP system, a CRM, and several spreadsheets. Copy jobs pull from each source, Dataflow Gen2 standardizes the formats, and the output loads directly into a Fabric Data Warehouse that Power BI reads from.
Copy jobs handle extraction from multiple source systems on a single schedule Dataflow Gen2 standardizes formats and applies cleansing logic before the data lands in the warehouse 2. Near Real-Time Analytics on Operational Data A retail operations team needs inventory and order data refreshed continuously, not in overnight batches. Mirroring replicates SQL Server data into OneLake in near real-time, giving analysts access to data that is minutes old rather than hours old.
Analysts work with near real-time data without any additional infrastructure to maintain Operations teams can act on current inventory and order status rather than yesterday’s snapshot 3. Migrating from Azure Data Factory An organization running Azure Data Factory pipelines needs to bring them into Fabric’s ecosystem to take advantage of the unified monitoring hub, Copilot, and native Lakehouse connectivity. Microsoft provides tooling for both the assessment and the migration itself.
PowerShell upgrade module handles pipeline conversion from ADF to Fabric Existing pipeline logic is preserved, and teams gain Fabric-native monitoring and Copilot on top
Kanerika’s FLIP-powered migration accelerator also automates the migrations ADF and synapse pipelines to Fabric, saving hours of manual work, costs and resources.
4. Self-Service Data Prep for Analysts An analytics team works with data from SharePoint, APIs, and flat files. They need to clean and reshape it regularly, but every change currently requires an engineering ticket and a wait. Dataflow Gen2 changes that dynamic.
Analysts build and maintain transformation logic through the visual Dataflow Gen2 interface Preview-only steps let them validate changes against sample data without touching production Engineers are freed from routine transformation requests and focus on higher-complexity pipeline work 5. Automating Business Workflows Alongside Data Pipelines A pipeline finishes loading data and the operations team needs to know . Rather than building a separate notification system, the pipeline includes an Office 365 Outlook activity that sends a configured email on completion.
Office 365 Outlook activity sends automated notifications when pipeline runs complete or fail Pipelines can update records in connected systems, turning a data job into a broader business workflow Already on Azure Data Factory? Here’s What to Know The use cases discussed above assume a clean start on Fabric. But for many organizations, the more pressing question is what happens to the Azure Data Factory investment they already have.
Fabric Data Factory is the next generation of ADF, built on the same foundation but with meaningful additions that ADF does not have. Here is how the key differences break down.
1. CI/CD Is Simpler in Fabric There is no dependency on ARM templates in Fabric. Teams can cherry-pick individual workspace items for check-in and check-out, and the built-in deployment pipelines work without needing an external Git repo.
No ARM template dependency for CI/CD workflows Built-in deployment pipelines work without an external Git repo Teams can cherry-pick individual pipeline items for version control rather than committing the full workspace 2. Monitoring Is More Useful The monitoring hub in Fabric provides cross-workspace visibility that ADF’s dashboard does not. Troubleshooting copy activities includes a detailed breakdown view that shows significantly more than the ADF equivalent.
Cross-workspace monitoring from a single hub, not per-pipeline dashboards Detailed copy activity breakdown for faster troubleshooting Copilot provides error summaries with root cause analysis and suggested fixes directly in the UI
3. ADF and Synapse Pipelines Can Run from Inside Fabric The new pipeline activity lets teams run and monitor existing ADF and Synapse pipelines from within Fabric, alongside Fabric dataflows and notebooks. This makes it practical to start using Fabric before committing to a full migration.
Run and monitor ADF and Synapse pipelines natively from a Fabric workspace Combine existing ADF pipelines with Fabric dataflows and notebooks in the same workflow Evaluate Fabric capabilities in parallel with existing ADF workloads before committing to a full cutover 4. Connector Parity Is Still Catching Up Most connectors are available in Fabric Data Factory, but organizations with specialized or less common sources should check the connector continuity documentation before setting a migration timeline.
Most major connectors are available and in parity with ADF Specialized or legacy connectors may not yet be fully supported in Fabric Microsoft’s connector parity documentation is the most reliable reference before planning a migration
For teams ready to move, Microsoft provides a PowerShell upgrade module and an ADF-to-Fabric item migration feature, both documented in the Fabric migration guide.
Getting the migration right is where working with an experienced implementation partner makes a meaningful difference.
How Kanerika Helps Organizations Implement and Migrate to Microsoft Fabric Data Factory The value of Fabric Data Factory depends heavily on how it is set up. Connecting 170+ data sources is possible, but knowing which connections to build first, how to structure pipelines for long-term maintainability, and how to configure governance correctly takes experience with the platform.
Kanerika is a Microsoft Solutions Partner for Data and AI and a Microsoft Fabric Featured Partner. The team includes Fabric-certified engineers, MVPs, and Superusers who work with organizations at every stage of Fabric adoption , from initial architecture to full platform rollout.
Microsoft Fabric Implementation Kanerika helps organizations design and build Fabric environments where Data Factory is one piece of a larger, integrated data architecture. That includes OneLake structure, data warehouse design , governance configuration, and pipeline strategy across all connected workloads.
Implementation is not treated as a standalone project. The goal is an architecture that supports the organization’s data needs today and scales without a rebuild as those needs grow.
For organizations moving off Azure Data Factory, Synapse, or on-premises ETL tools, Kanerika’s FLIP migration accelerator platform reduces the time and risk involved in the transition.
FLIP migration paths relevant to Data Factory:
Azure to Fabric Migration: moves existing Azure data estates into Fabric’s native environment
The Azure to Fabric Migration Accelerator reached General Availability as a Microsoft Fabric workload , making it a production-ready, validated path for organizations ready to move.
Case Study: Achieving a 90% Data Accuracy for a US Material Handling Company with Fabric across a wide network of service centers and warehouses across the US. Before working with Kanerika, their data was fragmented and unreliable.
The challenges they were dealing with: Data across SQL Server and SharePoint was completely siloed, with no central repository Inconsistent data quality was skewing KPI reporting across operational teams No unified architecture meant real-time decision-making was not possible What Kanerika built: A Data Lakehouse on Microsoft Fabric integrating data from SQL Server and SharePoint
A full data cleansing and validation process to fix KPI reliability at the source
A Power BI reporting framework with role-specific dashboards for different operational teams
The results: 90% improvement in data accuracy and KPI reliability 85% increase in operational visibility across service centers 100% scalable architecture built to grow with the business
A fragmented data estate became a clean foundation for operational decisions, without a multi-year transformation project.
A fragmented data estate became a clean foundation for operational decisions, without a multi-year transformation project.
Wrapping Up Upgrading your data infrastructure is a significant decision, and Microsoft Fabric makes a strong case as the platform to build on. Data Factory sits at the center of that, giving teams a single place to connect, move, transform, and orchestrate data across an entire estate, with AI assistance, unified governance, and direct integration with the rest of the Fabric platform built in.
Getting there faster, with less risk, is where the right partner matters. Kanerika brings certified expertise, production-tested migration tooling through FLIP, and a team that has done this across industries. Talk to us today to get started.
Transform Your Data Analytics with Microsoft Fabric! Partner with Kanerika for Expert Fabric implementation Services
Book a Meeting
FAQs No FAQ found.
