calendar 15, June 2026
by: Arreoblue

In Conversation with Dirk, BlakYaks: Scaling securely and building for the future

Dirk Anderson doesn’t oversell AI, but he doesn’t dismiss it either. As CEO of BlakYaks, a Microsoft-focused partner specialising in Azure platform engineering, he’s spent enough time inside large organisations to know the gap between what gets talked about and what actually gets built. In this episode of the Arreoblue podcast, Stephen Crisp spoke with him about where most businesses really are with AI at the moment, and what it actually takes to turn a promising experiment into something that holds up in production.

Where most organisations actually are

Earlier than they think is probably the simplest answer. There’s a lot of AI activity going on inside organisations, but most of it sits in pockets. Teams are picking up tools, trying things out, solving real problems, but it rarely connects to anything governed or strategic. It just builds up over time, outside the platform, without anyone really owning it.

“Most businesses probably don’t fully appreciate how broadly AI tools are already being used inside their own organisations. You see serious work happening in data teams, developer teams, DevOps teams, but it rarely makes its way onto a properly governed platform. That’s still the exception rather than the norm.”

What tends to be missing isn’t the technology itself. It’s the link between what leadership says it wants to do and what the platform teams are actually building. Enterprise architects often sit right in the middle of that, but the connection doesn’t happen by itself. Someone has to make it part of the job.

From experiment to something that actually runs

The pattern Dirk described will sound familiar to most people. It usually starts with individuals using AI to make their own work easier. Some of those experiments turn into things the business starts to rely on, often without anyone planning for that to happen. That’s when the gaps in architecture and governance start to show.

“You start with lots of individual use cases, then some of those turn into critical business functions. At that point you need to think very differently about how they’re governed, how they’re managed, and how they’re supported.”

Left alone, that can quickly become what he called shadow IT at speed. Widely used, not always well understood, and difficult to untangle later. None of the questions are new, though. How’s the data protected? Who controls what gets deployed? How do you know what’s actually running? These are the same things enterprise teams have always had to think about. They just get skipped when everyone’s focused on getting something working.

Experimentation is fine, until it isn’t

Dirk isn’t against experimentation. Most useful ideas start there. The problem comes when there’s no clear path from experiment to production, and prototypes start piling up that nobody quite trusts to run day to day.

Different organisations approach this in different ways. Commercial businesses often push for speed, while the public sector tends to move more carefully, partly because the consequences of getting automation wrong are harder to contain. Either way, the need for solid foundations underneath doesn’t really change.

Building for the fact that everything will change

One thing Dirk kept coming back to was how quickly the technology is moving. Models change, tools change, vendors change, and platforms have to be built with that in mind. If the architecture is too tightly tied to one approach, it becomes hard to move later.

Flexibility isn’t something you add once everything is finished. It has to be there from the start, even if that feels slower at the time. It’s also part of why the partnership with Arreoblue works. Each side knows what it owns, and where the handoff sits. It sounds simple, but it makes a difference.

On building a business that knows what it is

Towards the end of the conversation, Dirk talked about building BlakYaks as a deliberately focused consultancy, and how easy it is early on to drift away from what you’re actually good at.

“You need to know exactly what you do and what you don’t do. It’s very easy at the start to take work that isn’t really your thing just to keep revenue coming in, but that can take you away from the market you actually want to build.”

He also came back to culture, not as something you write on a slide, but as something clients notice almost immediately.

“Customers can tell very quickly what kind of team you are. If you build the right culture inside the business, that’s what they see as well.”

Simple enough idea, but often the thing that matters most over time, whether you’re building a consultancy or building platforms that are supposed to last.

And in most cases, that’s still the part organisations are working through now.

If any of this sounds familiar, it’s a conversation we’re having with a lot of teams at the moment. Our team at Arreoblue are always happy to talk through where you are and what the next step looks like.

 

Want to learn more:

Download the full article here.

Listen to the full conversation here.

Get in contact

Have a question or want to make an enquiry?

Contact us

Our technology partners