Cloud Solutions

Azure Virtual Desktop gives your team a full desktop from any device, with the data staying where you put it rather than scattering across laptops.

Most Microsoft 365 tenants run on defaults nobody revisited. Entra ID, Conditional Access, and the security baseline are where the value is, and they need deciding rather than accepting.

The same is true of what you are paying for it. Microsoft 365 licensing accumulates rather than gets chosen, and most tenants are carrying seats nobody signs into alongside a plan mix that no longer matches who does what.

Every other decision in a Microsoft cloud environment rests on identity, and it is the part most commonly left at whatever the migration produced.

Who exists, what they can reach, which devices are allowed, what happens when someone signs in from somewhere unusual — those are configuration decisions in Entra ID and Conditional Access, and they are the difference between a tenant that is defensible and one that merely works. Getting them right afterwards means revisiting everything built on top.

It is also the piece that determines whether the security side has anything to work with, because almost every modern incident is an identity event before it is anything else.

Azure bills for what is running, which is a genuine advantage and the source of nearly every unpleasant surprise.

The recurring causes are unremarkable: environments sized against the old server's specification rather than actual use, test resources spun up for a project and never switched off, virtual desktops left running overnight and at weekends, storage kept on expensive tiers because nobody chose a cheaper one, and no reserved capacity bought for the workloads that demonstrably run all year.

None of that is a flaw in Azure. It is what happens when consumption billing meets an environment nobody owns. Cost management is an ongoing job rather than a setting, and it is one of the clearer arguments for having somebody responsible for it — Azure Virtual Desktop in particular trades capital cost for running cost, and the saving depends entirely on hosts shutting down when nobody is using them.

The instinct is to lift what exists onto cloud infrastructure and change nothing. It is the fastest route and sometimes the right one, and it carries an obvious cost: you have relocated your existing problems and started paying rent on them.

The more useful question is which applications belong on infrastructure at all. Some should become a service you consume rather than a server you run. Some should be retired — there is almost always something still running because switching it off felt risky in 2019. Some genuinely need to stay where they are, and something old with a hardware dependency or a per-machine license is worth identifying before a design exists rather than during the cutover.

Hybrid is a legitimate destination, not a failure to finish. What is worth avoiding is arriving there by accident and calling it a strategy.

The individual moves are scoped as projects — on-premises to Microsoft 365, Google Workspace across, databases into Azure — because each depends on the state of what you have now.