Enterprise AI: the key signals from FabCon 2026

Some conferences confirm what you already know. Others change the way you see things. FabCon 2026 was clearly the second kind. Without fanfare or hype, the same observation kept coming up everywhere: enterprise AI cannot be built without truly usable data. Data that can flow in real time, unified, governed and contextualised at enterprise scale.

What seemed difficult to achieve just a few months ago is starting to take concrete shape today, thanks to the evolution of Microsoft Fabric and other major platforms in the market.

 

The foundations are set

The sessions highlighted a reality that many organisations know well: AI only works if the data it relies on is ready. Not just accessible, but coherent, reliable and structured for AI use cases. The announcements around Microsoft Fabric gave the sense that this goal is now genuinely within reach. What struck me was the speed at which Fabric is evolving. This is not a gradual evolution. It is an acceleration. Microsoft’s investment is clear, consistent, and visible in every announcement.

Fabric IQ now integrates semantics, ontologies and agents at the core of the platform. Data Agents reach general availability and integrate natively into the M365 Copilot experience. OneLake governance and security also reach GA. SAP Datasphere mirroring, combined with bidirectional interoperability with Snowflake and Databricks, is breaking down barriers that have fragmented enterprises for years.

What personally stood out to me: the platform is now genuinely ready to help organisations build their AI foundations, not as a future promise, but as something actionable right now. A business user can today create their own agents. The Copilot ecosystem simplifies the development of data pipelines and data models. A platform that accelerates AI development and uses AI itself to keep improving, continuously enriching the data layer going forward. Twelve to eighteen months ago, Fabric was not yet at this level. Today it is.

These are not incremental updates. They are structural shifts. Fabric is evolving from an analytics platform into a true operating system for enterprise AI. And other platforms like Databricks are following the same trajectory.

For Hitachi Solutions, this is concrete. These new capabilities give us the means to help our clients simplify their data environment, strengthen governance, unify their platforms and build architectures that can evolve at the pace of AI.

Art Cabugason from CDW shared a pragmatic, grounded approach to structuring Fabric capacities within an enterprise. Daniel Retzer from Hitachi Energy presented how his team is modernising an entire suite of energy network solutions using Fabric, Foundry and an agentic architecture, a concrete example of transformation already in motion. LexisNexis demonstrated a structured, repeatable method for building and testing agents via the Fabric SDK: notebooks, Q&A pairs, iterative refinements through Copilot.

What we saw was not demonstrations. It was real, already in production.

Then came Operations Agents in public preview. A quiet announcement, but with real implications. They mark the shift from reactive workflows to proactive, semi-autonomous operations in near real time. AI is no longer just a tool you call upon, it is starting to actively contribute to the running of operations.

The session that stayed with me the most was the Power BI Turing Test. An experienced Power BI expert and a novice are competing on a series of tasks. On some tasks, the expert was clearly ahead. On others, the results were almost equivalent. AI can genuinely accelerate Power BI development, but expertise remains essential for shaping the design, assessing quality, and identifying and correcting errors.

This is not a replacement. It is amplification. And that is exactly how we think about AI at Hitachi Solutions.

What does this mean for Hitachi Solutions and our clients?

After this conference, one thing is clear: the industry is no longer experimenting. It is building. And Fabric is becoming the backbone of that acceleration, alongside other platforms like Databricks.

Fabric’s governance capabilities, its agentic foundations and its architecture give us the means to help our clients move from isolated innovation to enterprise-wide transformation.

Modernising existing systems, building on approaches that have already proven their worth. Building AI-ready architectures that scale with confidence. Guiding organisations through the complexities of adoption with solutions that are secure, governed and operationally sound.

What I also take away is that the growing maturity of analytics and AI capabilities, driven by Fabric but also by other major platforms, highlights what we still need to build in terms of skills, methods and delivery quality. That is a clear commitment on our part.

FabCon 2026 confirmed where Microsoft Fabric is headed. And more importantly, what our clients can do with it and how we can help them get there.