Hiya! FabCon Atlanta 2026 has just wrapped up (March 16–20), and I am still buzzing. If you weren’t there, the one thing I’d encourage you to do is watch the keynote recording — “Unifying the Data Estate for the Next AI Frontier.” Nearly two hours, worth every minute.
The Central Idea
We’ve always been working towards a central data estate with decentralised consumers — the idea that data lives in one trusted, governed place, and different teams, tools, and users plug into it on their own terms. That’s not a new goal. But now AI is infused into the data space, and it completely changes the question we need to be asking: what does a central data estate actually look like when AI is part of the picture?
The keynote walks through exactly that — and it covers improvements across both Microsoft Fabric and Azure Data, which I loved. This isn’t just a Fabric story, it’s a broader platform story. Think about what AI participation in your data estate could look like:
- AI agents as coworkers — completing tasks, checking in, handing off work autonomously
- AI that can answer questions directly of your data, without someone having to write a query
- AI that creates notebooks, writes stored procedures, and accelerates the work of your data engineers and analysts
Real Customer Stories
Hearing executives from Coca-Cola and Delta share how they’re running on Fabric was honestly one of the highlights of the whole week. These aren’t small experiments — these are large-scale companies with genuinely complex, challenging data estates, and they’re making it work. Powerful stuff, and exactly the kind of business case you can point to when making the argument internally. And if neither of those resonate, Microsoft has a ton of customer stories across loads of their products and platforms — worth a browse at microsoft.com/en-us/customers. 🙌
RTI — Watch This Space
Real-Time Intelligence is an area that genuinely intrigues me — it’s a whole new way of thinking about data, and I find that equal parts exciting and challenging. If you want AI agents that can observe what’s happening in your business and act on it, they need data that’s current — not last night’s batch load. RTI as a first-class capability in Fabric is exactly the right direction. I’m genuinely thrilled to be spending more and more of my time going deeper and deeper on RTI capabilities — it’s an area I just can’t get enough of right now.
Agents as Coworkers
The framing I keep coming back to from the keynote: AI agents aren’t tools you pick up — they’re coworkers on your team. That shift in thinking matters because:
- Coworkers need context and shared understanding, not just instructions
- Coworkers can observe a situation and take initiative
- Coworkers need to understand what your data means — which is exactly why a governed semantic layer is so critical. This is where Fabric Ontology comes in, providing the shared context and meaning that lets agents (and humans!) actually understand your data, not just query it
Your agents are only as good as the data they have access to. Everything clicks when you see it through that lens.
Go Watch It
For those of us who’ve spent careers on data quality, governance, and integration — our moment is here. Those problems are now directly in the critical path for AI.
