Azure Virtual Desktop and Citrix both deliver Windows desktops and apps. They are not interchangeable once you have profile management, GPU needs, printer mapping, and a service desk that already knows one of the stacks.
AVD is the shorter path if you are already committed to Azure identity, want Microsoft to run the control plane, and can live within FSLogix, host-pool design and Azure networking. It is cheaper to start. It becomes expensive if you ignore image hygiene, autoscale and storage for profiles.
On one desktop project, a finance team ran a macro-heavy Excel workbook that read from a legacy ODBC source. It opened in seconds on the pilot AVD pool, but took a minute in production because the FSLogix profiles were on the wrong storage tier, and users blamed the platform. Moving the profile storage fixed it. Profile placement and image contents decided the experience more than the choice of broker.
Citrix still earns its keep for complex application publishing, certain WAN conditions, and organisations that have deep existing operational muscle. You pay for that control plane and for people who can run it. “Moving to AVD to save licences” fails when the hidden cost is a redesign of profiles, printing and application packaging.
Decide from constraints: multi-session density, GPU, partner access, sovereign or disconnected requirements, and who will patch images at 2am. Run a proof on the hardest user cohort, not the easiest. Finance users with macros and line-of-business apps will tell you the truth that a standard Windows desktop will not.
Hybrid is common. Some published apps stay on Citrix while task workers move to AVD. That is an architecture, not a failure, provided identity, profile and network paths are designed once rather than twice. Internal IT usually owns the user relationship; the engineering gap is the platform design.
Have a platform or landing zone project to get done? Transformation Engineering
All resources