Surfing the Semantic Wave












Scroll to Learn More ↓

Surfing the Semantic Wave








Scroll to Learn More ↓

Nicolas Castellvi

Consultant

Meet Nic


Nic is a data analytics and business intelligence consultant, passionate about turning complex challenges into simple, actionable insights. With strong experience across different industries, he focuses on building intuitive, user-friendly BI solutions that empower organisations to make smarter decisions. Often acting as the bridge between technical teams and business professionals, Nic enjoys helping users make sense of data and leverage it confidently. He is all about creating value, solving problems, and helping organisations grow through data-driven strategies.

Nicolas Castellvi

Consultant

Meet Nic


Nic is a data analytics and business intelligence consultant, passionate about turning complex challenges into simple, actionable insights. With strong experience across different industries, he focuses on building intuitive, user-friendly BI solutions that empower organisations to make smarter decisions. Often acting as the bridge between technical teams and business professionals, Nic enjoys helping users make sense of data and leverage it confidently. He is all about creating value, solving problems, and helping organisations grow through data-driven strategies.

If you told me five years ago that the hardest part of analytics would eventually be defining the meaning of “revenue,” “active customer,” or “cost per acquisition,” I might have laughed. But here we are.


I recently became the first person in Australia certified in Omni. Diving deep into its architecture, philosophy, and practical trade offs has convinced me of something: the semantic layer now sits at the crossroads of governance, flexibility, AI, and self service. Getting it right isn’t optional anymore: it’s foundational.


The whole point of BI has always been translation: taking data and using some abstract, geeky tools to turn it into definitions people can understand, trust, and use, without losing anyone along the way.



If you told me five years ago that the hardest part of analytics would eventually be defining the meaning of “revenue,” “active customer,” or “cost per acquisition,” I might have laughed. But here we are.


I recently became the first person in Australia certified in Omni. Diving deep into its architecture, philosophy, and practical trade offs has convinced me of something: the semantic layer now sits at the crossroads of governance, flexibility, AI, and self service. Getting it right isn’t optional anymore: it’s foundational.


The whole point of BI has always been translation: taking data and using some abstract, geeky tools to turn it into definitions people can understand, trust, and use, without losing anyone along the way

Know Your History to Know Where You’re Going

I’m a regular surfer, to be honest though, more of a casual one than a good one. Still, I’ve learned to appreciate the rhythm of waves: some you ride smoothly, others are hard to catch, and sooner or later a bigger set rolls in that changes everything.


BI has always moved in waves too, swinging between extremes. Early tools were tightly controlled, giving consistency but very little flexibility. Then empowering platforms like Power BI and Tableau flipped the script: anyone could explore and build, but the single source of truth was somehow lost, questioning the trust in the numbers. This was when we saw another generation of tools bring governance back into the conversation by introducing a semantic model layer. Big step forward, but still too complex for many users.


This has been the BI dilemma for years:

  • How much freedom do you give?
  • How much control do you keep?
  • How do you stop this from becoming a permanent trade off in the first place?

From Dashboards to Conversations

Then came AI, and the rules changed overnight. Users now expect to type a question into a chat window and get a reliable answer. No SQL, no hunting through dashboards, no waiting for a data team. The “report factory” model of BI, where engineers push dashboards to the business, suddenly feels dated.


That means business logic buried in ETL or scattered across tools doesn’t work anymore. We need a shared foundational layer that can serve AI, dashboards, and human exploration equally well.

Sometimes It’s Better to Dump It All and Start from Scratch

Now, there are different types of boards depending on the wave you need to surf. You might still love that old board as it’s familiar, you know how it behaves, but when the swell changes, it just can’t keep up. You can force it, sure, but you’ll either wipe out halfway or miss the full potential of the wave.


That’s exactly where BI is right now: trying to surf the AI wave with tools that were never shaped for it. The reality is most major BI platforms were built a while ago, and have been acquiring other products or stacking new features over time, often resulting in a patchwork of capabilities that don’t quite fit together and certainly don’t feel like a great user experience.


With AI now reshaping expectations, it might be time to rethink BI from the ground up: fresh, with the semantic layer at the core. So the real question becomes: do we keep patching the old board, or is it time to shape a new one?

What’s a Semantic Layer Anyway?

At its core, the semantic layer is about translation. It turns technical concepts into definitions that everyone understands. Tools like Omni even allow development flow in both directions; engineers will define logic centrally for everyone, but users can also experiment and propose their definitions to be added into shared models.


Of course, this only works with governance. Without guardrails, infinite truths will return. That’s why the semantic layer is so critical: it’s the one place where experimentation and trust meet. With everyone trying to use AI, the semantic layer will be the context that helps BI teams deliver meaningful, actionable insights instead of random guesses.


The space is moving fast. Industry wide efforts like the Open Semantic Interchange (OSI), led by Snowflake, are gaining traction. OSI is a vendor free standard: define your metrics, dimensions, and business logic once, and any compatible tool can use the same definitions.

Just Surf The Wave

Great surfers adapt, they don’t fight the wave. That’s the mindset for BI now. The landscape is shifting, but in essence the goal remains the same: helping people make informed decisions with data. The difference is we can do it smarter, faster, and more intuitively.


Developers - shape your development lifecycle around feedback and iteration. Governance should enable and encourage contribution, instead of bureaucracy. Stay curious about new tools and architectures, even before they’re part of your day-to-day.


Business Users - you’re no longer just consumers, you’re co-creators. Engage with how data is defined, and you’ll become more empowered, self-sufficient, and impactful.


Decision Makers - investing in semantics is high-leverage. If you want to encourage AI adoption, data-driven decision making, self serve analytics and governance, then get the semantics right, and set your team up for success.


Get ready, a big wave set is coming!