Dynamics 365 on prem is not dead: Insights from Philip Verlinden, Solution Architect

In this Elemental Talks episode, Philip Verlinden, Solution Architect at Business Elements Reply, challenges the widespread assumption that Dynamics 365 Customer Engagement on-premises is a thing of the past. Drawing on his experience with regulated industries such as banking and the public sector, Philip explains why on-prem remains a relevant and often necessary option in 2026, particularly in the face of strict data residency and compliance requirements. He explores what a modern on-prem Dynamics 365 setup looks like today, leveraging containers, Kubernetes, and even on-prem AI and shares concrete examples of how Business Elements Reply helps organizations modernize while preserving the power of choice between cloud and local environments.
Watch the video to gain a pragmatic, architecture-level perspective on the future of Dynamics 365 on-premises.
Interview:
There’s a perception that Dynamics 365 Customer Engagement on premises is obsolete. Is it the case?
Currently, today, we see in the industry that still a lot of different actors or organizations still need to have Dynamics 365 on premise. And it’s not only Dynamics. Of course, Dynamics itself didn’t have any recent large roadmap meaning like functional extensions or capabilities that you would have seen in the cloud version. But you still see that Microsoft is investing a lot in Azure local capabilities, but also Microsoft 365 local capabilities. And if you then need a business application and you have regulatory constraints or complaince frameworks needing you to kind of manage customer data on-premise, then Dynamics 365 is still a solid choice for organizations.
What are the main reasons why banks or public institutions still rely on Dynamics 365 on premises today?
In many cases, it’s due to the more local constraints or laws that they have to manage the data. For example, really, in the same custody of the country or the location itself, which makes sense. And certainly, that’s also where Microsoft still supports these customers. That’s why we still have Dynamics 365 customer engagement on-premise and also other related technology like Exchange, SharePoint, etc. But it all depends on the country and the local legislations and the frameworks that are in place. So, it depends and it’s still there. We still have the power of choice.
What does a “modern” Dynamics 365 on prem setup look like in 2026?
That’s a good question. If we think about it, then actually it’s not the same anymore like 15 years ago. 15 years ago, it was all on-premise. It was like that. Today, we also have more modern technologies available, like, for example, Docker and Kubernetes, which allow us also to run other workloads more easily on-premise. Or in, let’s say, a local data center. We can also extend capabilities that we have from Dynamics with also modern services or technologies. An example is that Microsoft also delivers AI services in containers. For example, text recognition, translation services, which you can run in your own data center on-premise in Docker or Kubernetes. But not only Microsoft, also other open-source capabilities are available. If you look at the Apache stack, for example, you’ll find a lot of containerized capabilities. And with that, you can do a fairly good mix of, capabilities and extend your own implementation in a very intelligent way, including using AI and large language models on-premise, for example. Of course, you will still need the CPU, the I .O. The infrastructure that needs to be in place, but you can do a lot on-premise as well from that perspective. So that’s, for me, certainly how a modern kind of setup would look like.
Can you share a practical example of how Business Elements Reply is helping regulated customers modernise or bridge to the cloud?
A practical example is indeed using those newer technologies like Machine Learning on-premise. And we can start small with the already known as large language models that exist and apply that also on data on-premise without sending data to the cloud. Another example is doing, for example, named entity recognition. That’s extracting data, specific data from documents. For that, we kind of learn or train the model with custom named entity recognition, even on-premise without sending data to the cloud, and then still have a benefit of using that technology on-premise without any cloud connection. And that’s an added value for a lot of our customers that still don’t have that possibility to go to the cloud even for management capabilities. Of course, that’s more for the air-gapped or the more restricted environments. If you have some possibilities to extend and use the best of both worlds; for example, have your data running on-premise but still have capabilities or management capabilities in the cloud like telemetry, then for me, that’s a win-win. Again, it’s the power of choice and you can cherry-pick what you would like to do.
What are the trade-offs or challenges that come with staying on prem versus moving to the cloud?
Also a very good question. What I see as a big difference is that in terms of management or operational capabilities, you need to do a little bit more on-premise. It’s not about opening an application, click “next, next, next” and everything is working. That’s not really the case. Also, introducing other capabilities and maybe other vendors or open-source capabilities will inflict also a broader, more diverse team with other backgrounds, other kind of people that have other experiences. And you will need to fine-tune the team working, the dynamics team, the tooling that you use, and find a common ground so that everyone is a little bit working in the same way. So, certainly there, we see some ripple effects that have some influence, but nevertheless, I think having more diverse teams also has a benefit. If you have diverse teams, you will learn from each other, and you’ll also accelerate later on during the implementation of a project. But that’s, I think, the biggest negative effect, but it also has positive side effects also from that perspective.
Finally, what’s your advice to CIOs or IT leaders deciding their next step?
Certainly do the right, correct assessment in terms of your data protection needs. Make sure you have everything lined up. And, indeed, if there is possibility to use cloud technology left or right in terms of operational perspective, also consider that. I think we are all aware that we need to stay or remain in certain compliances on-premise in terms of data, custodiality, the government in different European countries for example, which makes sense. It’s not only in Europe, also other countries do that, by the way, in the world. But then nevertheless, there are still possibilities outside of that, that you still can leverage. But again, first documents, get it first on paper. Black on white. And then take it from there and get somebody involved that knows how these things run. Do a call out to us to REPLY to Business Elements. We are there to help you.