If you were lucky enough to attend, reports are that It was a full day focused on how Microsoft Fabric and Azure databases work together to support real-time reporting and AI work for Australian businesses.
I do not think the timing was random.
This year, Fabric is continuing to gain momentum. More teams want one place to manage data, modernise older database and integration tools, and set themselves up for AI that is safe and repeatable.
Quick snapshot for leaders
If you are short on time, here is what to expect and why it matters.
Real-time intelligence: getting fresher data into dashboards and alerts so teams can act faster.
Unified data and governance: using OneLake, cataloguing, and clear ownership so self-service does not turn into chaos.
AI-ready data foundations: building reliable pipelines and governed workspaces so Copilot-style experiences can be trusted.
A practical migration path: reducing reliance on older integration patterns and moving toward Fabric Data Factory and modern lakehouse approaches.
In the rest of this blog, I will walk through the key Fabric changes we are seeing this year, what the public roadmap suggests is coming next, and how the roadshow topics translate into practical decisions for business leaders.
Why Fabric updates in 2026 matter
If you have been watching Fabric from the sidelines, the 2026 story is pretty clear: stronger governance, more AI features where they count, and more support for moving off older platforms like Azure Data Factory and on-premises data stacks.
For leaders, that usually comes down to three things:
lower integration cost
fewer data silos
a clearer path to getting value from AI without increasing risk or spend
Microsoft is also making it easier to keep up. Fabric changes are now published in a central “What’s new” experience, alongside a public roadmap. That helps me plan with clients because we can line up work with known release windows, rather than chasing announcements across blogs and conference sessions.
The big January 2026 wave
January delivered a solid wave of enhancements across Fabric. I do not see these as small nice extras.
Many of them reduce day-to-day friction for teams and improve how governance holds together at scale, which matched the themes discussed at the Brisbane roadshow.
Highlights
AI-powered catalogue and governance
The January updates included new AI experiences in the catalogue, designed to auto-summarise and better describe assets like semantic models. This helps non-technical stakeholders find what they need and understand it faster.
In practice, it also helps data teams keep self-service safer without needing a large team of data stewards. When people can discover data properly, they are less likely to create shadow copies and conflicting definitions.
OneLake security and hierarchy improvements
OneLake also received improvements, including parent-child hierarchy views in the catalogue and tighter security behaviour for mirrored databases. This supports a model where business domains can own their data areas, while central IT still sets guardrails.
If you want finance, operations, and marketing working from the same lake, these kinds of controls are what make that possible without teams stepping on each other’s toes.
Performance and reliability focus
There were also improvements aimed at performance and reliability, including proactive statistics refresh and other platform tuning.
For leaders, this matters because it supports sustained load. It becomes more realistic to run more reporting and AI workloads in Fabric without quickly hitting a performance limits.
That matters even more when the conversation is about real-time intelligence and AI scenarios that depend on fresh, fast data.
Data Factory in Fabric: migration and scale got real
One of the most important investment areas in 2026 is Fabric Data Factory. This is especially relevant for businesses still using Azure Data Factory, SSIS, or older on-premises integration tools.
Microsoft’s direction is clear: Fabric is where cloud data integration is heading, and the features landing now support that position.
If you attended the Brisbane roadshow, this topic was worth paying close attention to because modern data architecture is only as good as the pipelines that feed it.
Key changes
On-premises data gateway (January 2026 release)
A new gateway build (version 3000.302) shipped in January with performance improvements for reading CSV data in copy jobs and pipeline activities, plus adaptive tuning for read and write operations.
This is directly relevant if you are lifting data from file-based exports, line-of-business systems, or older SQL servers into Fabric. That is a common pattern when teams are bridging legacy systems with modern platforms.
Expanded incremental copy and change data capture (CDC)
Copy job now supports incremental copy from sources such as:
Google BigQuery
Google Cloud Storage
DB2
ODBC sources
Fabric Lakehouse tables
CDC options also expanded, including SAP Datasphere outbound paths to S3 and Google Cloud Storage.
For Australian businesses running hybrid or multi-cloud environments, this makes Fabric a more realistic integration hub, rather than a Microsoft-only island.
Secure orchestration with service principals and workspace identities
Recent updates allow Spark Job Definitions and notebooks to run in Data Factory pipelines using service principals or workspace identities instead of user credentials.
That is a big step toward production-grade automation that meets least-privilege expectations and can pass a security review.
Behind the scenes, Microsoft has also been publishing migration guidance and tools to assess an Azure Data Factory estate and plan a move to Fabric Data Factory.
If you are straddling multiple integration platforms, 2026 is a good time to put a consolidation plan on the table and validate it with Microsoft engineers and local partners at roadshow sessions like the one in Brisbane.
Platform enhancements leaders should care about
Alongside the plumbing, Fabric is adding platform capabilities that change what teams can do with data day to day. You do not need to memorise every feature name, but the themes matter, especially the ones linked to AI-ready data and real-time decision-making.
A few examples
AI functions across the stack
Fabric AI Functions became generally available in late 2025. They include capabilities like text embeddings, sentiment analysis, and structured extraction, and they continue to improve in 2026.
Practically, this means teams can build AI-supported classification, feedback analysis, or document tagging directly into Fabric pipelines and lakehouses, rather than wiring up separate services for every use case.
The roadshow agenda also talked about “agentic AI” scenarios. In plain terms, I think of agentic AI as AI that can take small steps across tools and data to complete a task, with guardrails.
AI Functions are part of what makes those scenarios possible without building everything from scratch.
Better support for complex, governed workspaces
Workspace-level configuration improvements and variable libraries that started landing in 2025 are maturing further in 2026. These help teams standardise business logic and parameters across projects.
This becomes important when you want hundreds of reports and pipelines to use the same definition of things like customer, margin, or region. Without that consistency, self-service turns into confusion fast.
Continuous integration and deployment (CI/CD) across Fabric
End-to-end CI/CD coverage for Fabric items continues to improve, including Git-backed development, better folder support, and stronger deployment pipelines.
For leaders, this influences release risk, auditability, and the ability to meet change control expectations, especially in regulated sectors.
Real-time intelligence: the roadshow’s centrepiece
If there is one theme that connected the Brisbane roadshow agenda with Fabric’s 2026 direction, it was real-time intelligence.
Microsoft is positioning Fabric as the platform where streaming data, operational databases, and analytics come together. The goal is simple: help teams act on data in minutes or seconds, not days.
This year’s updates support that in a few practical ways:
Closer integration between Azure databases (such as Cosmos DB, Azure SQL, and PostgreSQL) and OneLake, including the ability to mirror operational data without complex ETL pipelines.
Better incremental copy and CDC options that make it practical to keep Fabric aligned with transactional systems in near real time.
A clearer platform story that combines streaming, lakehouse, and reporting in one place, so teams spend less time moving data between tools.
Looking ahead: what the 2026 roadmap signals
Microsoft’s public roadmap and recent update patterns suggest a few clear priorities for 2026. I use these as planning cues when I talk with business and data leaders, because they point to where Microsoft is investing the most effort.
Publicly discussed roadmap themes include
More governance and cataloguing improvements, so scale does not come at the cost of control.
More integration support and migration tooling, especially for teams moving from older data integration approaches.
More real-time capabilities and better ways to work with operational data alongside analytics.
More built-in AI support across workflows, with clearer guardrails and identity-based controls.
What this means for Australian businesses
In Australia, I often see the same realities across mid-market, government, and not-for-profit teams:
hybrid data
older SQL estates
spreadsheet-heavy reporting
mixed sources across cloud platforms
That is why the combination of Fabric and Azure databases is interesting. It creates a path where you can modernise in steps. You can keep the systems that still need to run day to day, while you improve how data is collected, governed, and used for reporting and AI.
If you attended the Brisbane roadshow, the best value came from treating it like a chance to ask practical questions about your current environment, not just listening to product updates.
How I approach 2026 planning with Fabric
When I plan Fabric work with clients, I try to keep it simple and staged. Here is an approach that keeps the detail, but makes it easier to execute.
1) Start with governance basics that people will actually use
Decide who owns each data domain and what “done” looks like for documentation and quality.
Set access rules early, including least-privilege patterns for pipelines and automation.
Agree on naming and catalogue standards so data discovery works for business users.
2) Build a migration plan for integration, not just reporting
Identify your highest-value pipelines and assess whether they should move to Fabric Data Factory first.
Look for duplication across tools, including repeated extracts and manual spreadsheet steps.
Prioritise patterns that reduce support overhead and make audits easier.
3) Make real-time a business decision, not a tech hobby
Pick one or two decisions that would improve with fresher data, such as operational delays, customer service queues, or stock movement.
Work backwards to the data sources and CDC needs.
Define what “near real time” means for your business, because not every use case needs seconds.
4) Treat AI as a layer that sits on trusted foundations
Choose a clear pilot use case, such as document tagging, feedback classification, or policy search.
Ensure the data is governed, access is controlled, and outputs can be explained.
Measure success using simple metrics like time saved, fewer manual steps, or faster reporting cycles.
Questions worth taking to sessions like this
Even though the Brisbane roadshow has now wrapped up, these are still useful questions to take into future Microsoft sessions, partner workshops, or internal planning meetings. They help turn a conference day into a real plan.
What are the recommended patterns for mirroring Azure database data into OneLake, and what are the common pitfalls?
What migration tooling is available for Azure Data Factory, and what is the realistic effort for a mid-sized environment?
How should we structure workspaces and domains so business teams can own data while IT keeps consistent guardrails?
Which CI/CD options are recommended today for Fabric items, and what should we avoid in regulated environments?
How is Microsoft defining and supporting agentic AI scenarios in Fabric, and what controls are expected?
Final thoughts
I see 2026 as a year where many Australian businesses will move from Fabric curiosity to Fabric planning. The Brisbane roadshow was a useful moment to sense-check what Microsoft is recommending, and how that lines up with your current data and database direction.
At CG TECH, we help teams turn platform updates into a clear 90-day plan. That might be a governance check, a migration roadmap for integration, or a focused pilot that proves value without creating risk.
If you attended the roadshow and want a sounding board, I am always happy to compare notes.
About the Author
Carlos Garcia is the Founder and Managing Director of CG TECH, where he leads enterprise digital transformation projects across Australia.
With deep experience in business process automation, Microsoft 365, and AI-powered workplace solutions, Carlos has helped businesses in government, healthcare, and enterprise sectors streamline workflows and improve efficiency.
He holds Microsoft certifications in Power Platform and Azure and regularly shares practical guidance on Copilot readiness, data strategy, and AI adoption.
On 16 February 2026, Microsoft brought the Microsoft Fabric and Database Roadshow to Brisbane.
If you were lucky enough to attend, reports are that It was a full day focused on how Microsoft Fabric and Azure databases work together to support real-time reporting and AI work for Australian businesses.
I do not think the timing was random.
This year, Fabric is continuing to gain momentum. More teams want one place to manage data, modernise older database and integration tools, and set themselves up for AI that is safe and repeatable.
Quick snapshot for leaders
If you are short on time, here is what to expect and why it matters.
In the rest of this blog, I will walk through the key Fabric changes we are seeing this year, what the public roadmap suggests is coming next, and how the roadshow topics translate into practical decisions for business leaders.
Why Fabric updates in 2026 matter
If you have been watching Fabric from the sidelines, the 2026 story is pretty clear: stronger governance, more AI features where they count, and more support for moving off older platforms like Azure Data Factory and on-premises data stacks.
For leaders, that usually comes down to three things:
Microsoft is also making it easier to keep up. Fabric changes are now published in a central “What’s new” experience, alongside a public roadmap. That helps me plan with clients because we can line up work with known release windows, rather than chasing announcements across blogs and conference sessions.
The big January 2026 wave
January delivered a solid wave of enhancements across Fabric. I do not see these as small nice extras.
Many of them reduce day-to-day friction for teams and improve how governance holds together at scale, which matched the themes discussed at the Brisbane roadshow.
Highlights
AI-powered catalogue and governance
The January updates included new AI experiences in the catalogue, designed to auto-summarise and better describe assets like semantic models. This helps non-technical stakeholders find what they need and understand it faster.
In practice, it also helps data teams keep self-service safer without needing a large team of data stewards. When people can discover data properly, they are less likely to create shadow copies and conflicting definitions.
OneLake security and hierarchy improvements
OneLake also received improvements, including parent-child hierarchy views in the catalogue and tighter security behaviour for mirrored databases. This supports a model where business domains can own their data areas, while central IT still sets guardrails.
If you want finance, operations, and marketing working from the same lake, these kinds of controls are what make that possible without teams stepping on each other’s toes.
Performance and reliability focus
There were also improvements aimed at performance and reliability, including proactive statistics refresh and other platform tuning.
For leaders, this matters because it supports sustained load. It becomes more realistic to run more reporting and AI workloads in Fabric without quickly hitting a performance limits.
That matters even more when the conversation is about real-time intelligence and AI scenarios that depend on fresh, fast data.
Data Factory in Fabric: migration and scale got real
One of the most important investment areas in 2026 is Fabric Data Factory. This is especially relevant for businesses still using Azure Data Factory, SSIS, or older on-premises integration tools.
Microsoft’s direction is clear: Fabric is where cloud data integration is heading, and the features landing now support that position.
If you attended the Brisbane roadshow, this topic was worth paying close attention to because modern data architecture is only as good as the pipelines that feed it.
Key changes
On-premises data gateway (January 2026 release)
A new gateway build (version 3000.302) shipped in January with performance improvements for reading CSV data in copy jobs and pipeline activities, plus adaptive tuning for read and write operations.
This is directly relevant if you are lifting data from file-based exports, line-of-business systems, or older SQL servers into Fabric. That is a common pattern when teams are bridging legacy systems with modern platforms.
Expanded incremental copy and change data capture (CDC)
Copy job now supports incremental copy from sources such as:
CDC options also expanded, including SAP Datasphere outbound paths to S3 and Google Cloud Storage.
For Australian businesses running hybrid or multi-cloud environments, this makes Fabric a more realistic integration hub, rather than a Microsoft-only island.
Secure orchestration with service principals and workspace identities
Recent updates allow Spark Job Definitions and notebooks to run in Data Factory pipelines using service principals or workspace identities instead of user credentials.
That is a big step toward production-grade automation that meets least-privilege expectations and can pass a security review.
Behind the scenes, Microsoft has also been publishing migration guidance and tools to assess an Azure Data Factory estate and plan a move to Fabric Data Factory.
If you are straddling multiple integration platforms, 2026 is a good time to put a consolidation plan on the table and validate it with Microsoft engineers and local partners at roadshow sessions like the one in Brisbane.
Platform enhancements leaders should care about
Alongside the plumbing, Fabric is adding platform capabilities that change what teams can do with data day to day. You do not need to memorise every feature name, but the themes matter, especially the ones linked to AI-ready data and real-time decision-making.
A few examples
AI functions across the stack
Fabric AI Functions became generally available in late 2025. They include capabilities like text embeddings, sentiment analysis, and structured extraction, and they continue to improve in 2026.
Practically, this means teams can build AI-supported classification, feedback analysis, or document tagging directly into Fabric pipelines and lakehouses, rather than wiring up separate services for every use case.
The roadshow agenda also talked about “agentic AI” scenarios. In plain terms, I think of agentic AI as AI that can take small steps across tools and data to complete a task, with guardrails.
AI Functions are part of what makes those scenarios possible without building everything from scratch.
Better support for complex, governed workspaces
Workspace-level configuration improvements and variable libraries that started landing in 2025 are maturing further in 2026. These help teams standardise business logic and parameters across projects.
This becomes important when you want hundreds of reports and pipelines to use the same definition of things like customer, margin, or region. Without that consistency, self-service turns into confusion fast.
Continuous integration and deployment (CI/CD) across Fabric
End-to-end CI/CD coverage for Fabric items continues to improve, including Git-backed development, better folder support, and stronger deployment pipelines.
For leaders, this influences release risk, auditability, and the ability to meet change control expectations, especially in regulated sectors.
Real-time intelligence: the roadshow’s centrepiece
If there is one theme that connected the Brisbane roadshow agenda with Fabric’s 2026 direction, it was real-time intelligence.
Microsoft is positioning Fabric as the platform where streaming data, operational databases, and analytics come together. The goal is simple: help teams act on data in minutes or seconds, not days.
This year’s updates support that in a few practical ways:
Looking ahead: what the 2026 roadmap signals
Microsoft’s public roadmap and recent update patterns suggest a few clear priorities for 2026. I use these as planning cues when I talk with business and data leaders, because they point to where Microsoft is investing the most effort.
Publicly discussed roadmap themes include
What this means for Australian businesses
In Australia, I often see the same realities across mid-market, government, and not-for-profit teams:
That is why the combination of Fabric and Azure databases is interesting. It creates a path where you can modernise in steps. You can keep the systems that still need to run day to day, while you improve how data is collected, governed, and used for reporting and AI.
If you attended the Brisbane roadshow, the best value came from treating it like a chance to ask practical questions about your current environment, not just listening to product updates.
How I approach 2026 planning with Fabric
When I plan Fabric work with clients, I try to keep it simple and staged. Here is an approach that keeps the detail, but makes it easier to execute.
1) Start with governance basics that people will actually use
2) Build a migration plan for integration, not just reporting
3) Make real-time a business decision, not a tech hobby
4) Treat AI as a layer that sits on trusted foundations
Questions worth taking to sessions like this
Even though the Brisbane roadshow has now wrapped up, these are still useful questions to take into future Microsoft sessions, partner workshops, or internal planning meetings. They help turn a conference day into a real plan.
Final thoughts
I see 2026 as a year where many Australian businesses will move from Fabric curiosity to Fabric planning. The Brisbane roadshow was a useful moment to sense-check what Microsoft is recommending, and how that lines up with your current data and database direction.
At CG TECH, we help teams turn platform updates into a clear 90-day plan. That might be a governance check, a migration roadmap for integration, or a focused pilot that proves value without creating risk.
If you attended the roadshow and want a sounding board, I am always happy to compare notes.
About the Author
Carlos Garcia is the Founder and Managing Director of CG TECH, where he leads enterprise digital transformation projects across Australia.
With deep experience in business process automation, Microsoft 365, and AI-powered workplace solutions, Carlos has helped businesses in government, healthcare, and enterprise sectors streamline workflows and improve efficiency.
He holds Microsoft certifications in Power Platform and Azure and regularly shares practical guidance on Copilot readiness, data strategy, and AI adoption.
Sources and useful links
Glossary of technical terms (plain English)
Agentic AI
AI that can take small steps to complete a task across tools or data, using rules and approvals you set.
Azure
Microsoft’s cloud platform for hosting apps, databases, and services.
Azure SQL
A Microsoft cloud database service used to store and manage business data.
Catalogue (data catalogue)
A searchable library of data that explains what it is, who owns it, and how it should be used.
CDC (Change Data Capture)
Copies only what changed in a database (new or updated records), rather than copying everything each time.
CI/CD (Continuous Integration / Continuous Deployment)
A structured way to test and release changes safely using repeatable steps.
Data pipeline
Automated steps that collect, clean, and move data from one system to another.
ETL (Extract, Transform, Load)
A common approach to data work: pull data out, change it into the right format, then load it for reporting or analysis.
Fabric (Microsoft Fabric)
Microsoft’s platform that brings together data integration, storage, analytics, and governance in one place.
Fabric Data Factory
The Fabric tool used to build pipelines and move data between systems.
Git
A system that tracks changes to files or code over time, helping teams work safely and roll back if needed.
Incremental copy
Moves only new or changed data since the last run, which is faster and cheaper than full copies.
Lakehouse
A modern data setup that combines low-cost storage (data lake) with structured reporting and analytics.
Mirroring (mirrored database)
Keeping a near real-time copy of an operational database somewhere else so it can be analysed without slowing the original system.
Multi-cloud
Using more than one cloud provider, such as Microsoft Azure and Google Cloud.
Notebook
A workspace where data teams run code and analysis (often using SQL or Python).
ODBC (Open Database Connectivity)
A standard method for connecting to many different kinds of databases.
OneLake
Fabric’s shared data storage layer designed to keep data in one place with consistent access controls.
On-premises data gateway
A secure connector that lets Fabric access data stored on your on-site servers.
Service principal
A non-human account used by apps or automation to access services securely, without using a person’s login.
Semantic model
A layer that defines business-friendly data terms and relationships so reports stay consistent.
Spark
A tool used to process large datasets, commonly used for data engineering and advanced analytics.
SSIS (SQL Server Integration Services)
An older Microsoft tool used for on-premises data integration and ETL processes.
Streaming data
Data that arrives continuously (like transactions or app activity) rather than in batches.
Workspace identity
An identity tied to a Fabric workspace that can be used for secure automation and access control.
Recent Posts
Popular Categories
Archives