Both platforms can use Azure—but the control plane differs
Citrix’s current Azure reference architecture describes Citrix DaaS using Citrix Cloud control-plane services with customer resource locations in Azure. Citrix can manage application and desktop delivery across multiple resource locations and support hybrid or multicloud patterns.
AVD uses the Microsoft AVD service as the brokering and control plane and exposes Azure-native host pools, application groups, scaling, identity, network, and monitoring integrations. The comparison therefore includes platform layering and operational responsibility, not simply the VM that hosts Windows.
| Decision area | Citrix DaaS on Azure | Azure Virtual Desktop |
|---|---|---|
| Control plane | Citrix Cloud and Citrix management layer | Microsoft AVD service |
| Hybrid/multicloud | Strong multi-resource-location model | Microsoft AVD and Azure-centric patterns, including supported Azure Local scenarios |
| User experience/protocol | Citrix HDX and Citrix ecosystem | Microsoft RDP, Windows App, RDP Shortpath |
| Application/image tooling | Citrix application and image-management ecosystem | Azure/Windows image patterns, FSLogix and application delivery methods |
| Operational stack | Citrix plus Azure plus Windows/EUC | Azure plus AVD plus Windows/EUC |
| Licensing | Citrix subscription plus Microsoft/Azure dependencies | Microsoft eligibility or access rights plus Azure consumption |
Decision rule: retain Citrix where its differentiated capabilities are actively used and operationally valuable; simplify to AVD where Citrix has become a mostly transparent layer over an Azure-hosted Windows service.
Inventory the Citrix features users and operators actually depend on
Do not make the decision from the product brochure. Inventory application publishing, desktops, delivery groups and catalogs, policies, HDX features, gateways, profile tools, application layering, printing, USB and peripheral redirection, monitoring, session recording, identity integrations, multi-site failover, automation, and help-desk workflows.
Classify each capability as business-critical, operationally useful, replaceable, unused, or unknown. Unknowns are migration risk. An old Citrix policy may represent an application workaround or compliance requirement that no one remembers. Rebuilding or deleting it without understanding the reason can create avoidable failure.
Migration starts with dependency discovery
A Citrix environment often contains years of policy and exception history. The safest migration does not attempt to recreate every setting in AVD. It identifies which policies still solve a real user or security requirement, then redesigns those outcomes using the target platform.
Citrix can remain valuable in hybrid and complex estates
Citrix DaaS is designed to manage desktops and applications across resource locations, including Azure and other environments. Organizations with on-premises workloads, multiple clouds, complex application delivery, specialized peripherals, or standardized Citrix operational tooling may value that abstraction.
The cost of keeping Citrix should be compared with the cost and risk of replacing those capabilities—not only the subscription price. If Citrix prevents the organization from operating three separate regional VDI control planes or supports a critical application feature that would require custom engineering in AVD, that value belongs in the decision.
AVD can reduce platform layers for Azure-centric organizations
For an organization already standardizing on Microsoft Entra, Intune, Defender, Azure networking, Azure Monitor, and Microsoft 365, AVD can reduce the number of control planes operators must understand. Native autoscale, FSLogix, Windows App, Private Link and RDP Shortpath options, and Microsoft management integrations cover many mainstream use cases.
Simplification has value when the extra Citrix layer is not delivering enough differentiated capability. But migration work still includes applications, profiles, policies, images, user assignments, monitoring, support, and rollback. Removing a control plane does not remove end-user-computing engineering.
User experience needs side-by-side validation
Display protocol performance depends on user location, endpoint, workload, multimedia, network, peripherals, and policy. Do not assume either Citrix HDX or Microsoft RDP will be universally better. Create a representative pilot with the same applications and user locations and measure responsiveness, Teams and media, printing, USB, graphics, file transfer, reconnect behavior, and help-desk diagnostics.
The business should define which user-experience differences matter. A technical preference that users do not notice may not justify a complex platform; a specialized workflow that breaks can justify retaining one. Include remote and branch locations that represent the real network rather than testing only from the datacenter or IT office.
Cost comparison needs all platform layers
| Cost category | Citrix path | AVD path |
|---|---|---|
| Desktop/app platform | Citrix subscription and licensing | AVD access entitlement requirements |
| Azure infrastructure | Session-host or resource-location VMs, storage, network | Session hosts, storage, network |
| Management tools | Citrix operations and monitoring plus Microsoft stack | Microsoft or Azure tools and any third-party tools |
| Profiles/images/apps | Citrix and Windows tooling plus operational labor | FSLogix, image and application operational labor |
| Migration | Potentially lower if retaining platform | Discovery, rebuild, pilot and wave migration |
| Support | Citrix plus Microsoft plus internal teams | Microsoft plus internal or partner teams |
If the business case assumes immediate license elimination, check contract renewal dates, coexistence periods, and rollback. Migrations commonly require both platforms to run during validation and user waves. Include that overlap in the financial model so the migration year is not compared unfairly with a steady-state target.
Profiles and user state need a migration plan
Citrix environments may use Citrix Profile Management, FSLogix, persistent desktops, folder redirection, or combinations. AVD commonly uses FSLogix. Profile data should not simply be copied without understanding versions, Office caches, application state, exclusions, and user data locations.
Decide which data should migrate, which should be regenerated, and which belongs in OneDrive, SharePoint, a file service, or another business data platform rather than the profile. Coexistence is especially important: if a user can work in both Citrix and AVD, two systems should not make conflicting changes to the same profile or application state without a tested design.
Applications are usually the critical path
Build an application inventory with owner, version, install method, dependencies, licensing, concurrency, peripheral needs, and current Citrix delivery method. Some applications can install into a shared AVD image; others need separate host pools or personal desktops; some may be better modernized or delivered as SaaS.
Treat “works in Citrix” as evidence that the application can be virtualized, not proof that it will work unchanged in a different session-host and policy model. Validate application authentication, data paths, drivers, browser components, licensing servers, printing, and update behavior in the target environment.
Coexistence can be a strategy, not a failure
A controlled migration can keep specialized users or applications on Citrix while moving mainstream personas to AVD. That reduces the blast radius and gives the organization evidence about which Citrix capabilities are truly required. The downside is a temporary—or sometimes permanent—dual operating model that must be intentionally supported.
Use retirement criteria for Citrix capabilities
For each Citrix-specific service or policy, define the AVD capability that replaces it, the validation evidence, and the last user or workload that depends on it. This creates an actual retirement plan instead of a vague assumption that licenses can be cancelled after the final user wave.
Retirement can include gateways, delivery components, monitoring, profile services, application layers, automation, service accounts, certificates, DNS records, firewall rules, and vendor integrations. Some components may remain for a retained workload, so decommissioning should follow dependency evidence rather than a target date alone.
Dual-platform support needs an end state
During coexistence, help-desk tools, user communications, monitoring, and escalation may need to support both environments. Name the date or condition when dual support ends. Otherwise temporary coexistence can become permanent complexity without a business decision.
Support staff should be able to identify which platform a user belongs to, where the authoritative profile or data lives, and which vendor or team owns the incident. User-facing documentation should avoid sending the same person through two different connection paths without explanation.
A practical migration sequence
- Inventory: users, apps, policies, catalogs, images, profiles, network, peripherals and support processes.
- Segment personas: mainstream candidates, specialized candidates, and blockers.
- Map capabilities: identify AVD equivalents, redesigns, and Citrix-only requirements.
- Build an AVD pilot: use representative apps, profiles and users.
- Run side-by-side validation: compare experience, operations, support and cost.
- Migrate low-risk waves: retain rollback access to Citrix for a defined period.
- Address exceptions: redesign or retain Citrix for justified workloads.
- Retire intentionally: remove Citrix components and licenses only after dependency validation.
Rollback criteria should be agreed before the wave
- Critical application workflow fails and no approved workaround exists.
- User experience falls below agreed qualitative thresholds for a persona.
- Profile or data migration creates corruption or material lost state.
- Required peripheral or printing workflow is not supportable.
- Security or monitoring controls cannot meet the target baseline.
- Help desk cannot diagnose or route common AVD incidents.
- The migration would require an unplanned licensing or infrastructure change.
When retaining Citrix can be the right answer
Retain or partially retain Citrix when differentiated capabilities, hybrid or multicloud control, specialized peripherals, policy depth, protocol behavior, existing automation, or operational expertise materially support business requirements. The roadmap should record the reason, not assume every retained workload is migration failure.
A mixed platform should still have governance. Define which user or application characteristics justify Citrix and which default to AVD. Revisit retained workloads at renewal or architecture milestones. This prevents a small set of exceptions from silently preserving the entire legacy operating model.
Leadership questions for the platform decision
- Which Citrix features are actually used by critical users today?
- Which of those have a tested AVD equivalent or redesigned outcome?
- What does the Citrix subscription buy us beyond Azure and Windows services?
- How long will both platforms need to run during migration?
- Which personas have the highest protocol, peripheral, or application risk?
- Can the support team operate AVD at the required level without Citrix tooling?
- What evidence must exist before Citrix components and licenses can be retired?
Compare the operating model, not only feature parity
A platform decision changes who does the work. In Citrix, administrators may rely on Citrix Studio or cloud management, Citrix monitoring, Citrix policy, machine catalogs, and existing automation. In AVD, those responsibilities shift toward Azure, AVD host pools, Microsoft monitoring, Intune or policy tooling, image automation, FSLogix, and Azure cost management. The organization should identify which team will own each target capability and whether that team has the required skills.
Operational simplification should be measurable. If moving to AVD removes a control plane but creates three new custom scripts and a manual image process, the environment may not actually be simpler. Conversely, if Citrix is retained only because teams are familiar with it, compare the training and process change required for AVD with the recurring cost and complexity of maintaining the existing stack.
Pilot evidence should include support effort
A technical pilot should record more than user experience. Track how long it takes to provision a host, publish an application, troubleshoot a profile, identify a failed session, apply an image update, change capacity, and collect diagnostic evidence. These tasks reveal whether the target operating model is easier or harder for the real support team.
Ask the help desk to participate in the pilot. Give them representative incidents and see whether they can route identity, profile, application, host, and network problems without relying on the Citrix tools being replaced. AVD adoption is not complete if every issue still requires the engineer who built the pilot.
Contract timing can change the migration sequence
Citrix subscription renewal, Microsoft licensing dates, datacenter contracts, hardware support, and application renewal cycles can affect the financially sensible migration window. A project that technically finishes after a renewal date may create another full contract period of overlap. Procurement should therefore be part of the migration roadmap, but commercial deadlines should not force unsafe user waves.
Build at least two timelines: the technically preferred sequence and the commercially constrained sequence. If they differ, leadership can decide whether to accept overlap cost, accelerate low-risk waves, negotiate contract terms, or retain a smaller licensed footprint. This is more transparent than promising a license saving before the migration dependency map is complete.
Data and security controls need equivalence, not identical settings
Citrix policies may control clipboard, drive mapping, printing, session recording, browser behavior, USB, and other data paths. AVD can implement many related controls through RDP properties, Intune, Group Policy, Conditional Access, endpoint controls, application settings, and Microsoft security services, but the setting names and enforcement points differ.
For each important security policy, document the control objective and evidence required in the target. Test the outcome with representative users. The goal is to preserve or improve the security requirement, not to prove that every Citrix policy has a one-to-one AVD checkbox.
Where BI Cloud Tech can help
BI Cloud Tech can combine Azure Virtual Desktop expertise, a Migration Readiness Assessment, and Migrations and Modernization to inventory Citrix dependencies, compare target capabilities, build an AVD pilot, and plan controlled migration waves. Citrix licensing and product support should be validated with the customer’s current Citrix agreements and vendor guidance.
A practical next step
Export the top 25 Citrix policies and application groups and ask one question for each: “What business or technical problem does this still solve?” Any setting with no answer should not automatically be recreated in AVD. Contact BI Cloud Tech to request a Citrix-to-AVD assessment.