How to Migrate Azure Analysis Services to Power BI Without Rebuilding Reports
Microsoft Fabric enables a direct model migration path from Azure Analysis Services to Power BI Premium — without rebuilding downstream reports
Microsoft has signalled the strategic direction clearly: Azure Analysis Services is not the destination. Power BI Premium — now unified under the Microsoft Fabric platform — is where enterprise semantic modelling is headed. Organisations that have invested years in building SSAS tabular models on Azure Analysis Services now have a defined, supported migration path that preserves their existing Power BI reports, dashboards, and connections. The question is no longer whether to migrate Azure Analysis Services to Power BI, but how to do it with minimal disruption and maximum governance.
This guide is written for data leaders — CDOs, IT Directors, and Analytics Managers — who are responsible for making that call. It covers the full migration process, the prerequisites that must be satisfied before a single model moves, and the strategic considerations that determine whether a phased or full-cutover approach is right for your organisation.
Why Organisations Are Moving Off Azure Analysis Services
Azure Analysis Services was the right platform for enterprise tabular modelling throughout the 2010s. It provided high-performance in-memory analytics, enterprise-scale RLS, and robust connectivity to Power BI. For many organisations, it remains in production today supporting significant analytical workloads.
The case for migration has strengthened considerably over the past two years. Microsoft has consolidated its analytical investments inside Microsoft Fabric, meaning that new capabilities — Direct Lake mode, OneLake integration, Fabric Git source control, and advanced deployment pipeline features — are being built for Power BI Premium semantic models, not for Azure Analysis Services. Staying on AAS means accepting a growing feature gap as the platform matures around Power BI.
Beyond feature parity, the economics are shifting. Running a dedicated AAS server adds infrastructure cost that Power BI Premium capacity — shared across Fabric workloads — can often absorb more efficiently. For organisations already on Fabric capacity, migrating AAS workloads means consolidating spend rather than running parallel platform costs.
The Executive Business Case for AAS to Power BI Migration
The migrate Azure Analysis Services to Power BI decision is ultimately a platform consolidation play. Executives evaluating the case should frame it around three outcomes: cost reduction, capability unlock, and risk reduction.
On cost, consolidating AAS workloads into existing Fabric capacity eliminates a separate Azure resource that requires independent scaling, monitoring, and backup management. On capability, Power BI Premium semantic models now support Direct Lake mode — eliminating data import latency — and integrate natively with Fabric pipelines, Git workflows, and OneLake storage. On risk, Azure Analysis Services carries an increasing platform risk as Microsoft's investment flows almost entirely into Fabric. The earlier the migration, the lower the transition cost and the longer the runway to take advantage of Fabric's expanding capability set.
Critically, the migration process is designed to preserve existing Power BI report connections. Reports built against an AAS model do not need to be rebuilt — they are rebound to the migrated semantic model in Power BI Premium. This removes the single biggest objection to migration: the fear of regression in a report estate that has taken years to build.
Prerequisites: What Must Be in Place Before You Start
A failed migration is almost always a prerequisites failure. Before initiating any AAS to Power BI migration, confirm that every item in both checklists below is satisfied. Attempting to migrate without meeting these requirements will result in errors mid-process — at a point where partial migration state can create additional remediation work.
Azure Analysis Services Requirements
- The AAS server and the target Power BI workspace must reside in the same Azure tenant.
- The migrating user must have Server Administrator permissions on the AAS instance and hold Owner or Contributor roles on the Azure subscription.
- AAS must have an Azure Storage account configured with backup enabled. The migration process uses this backup storage to transfer model data.
- If a Firewall is enabled on your AAS server, the "Allow access from the Power BI Service" setting must be set to On — or the firewall must be disabled for the duration of the migration.
- The AAS server must be in a started (running) state during migration. It can be paused after migration is confirmed successful.
Power BI / Fabric Requirements
- The target workspace must be backed by a Fabric Capacity, Power BI Premium per Capacity, Power BI Premium per User, or Power BI Embedded licence. Shared capacity is not supported.
- The migrating user must have Workspace Administrator permission on the target workspace.
- An Azure Data Lake Storage Gen 2 (ADLS Gen 2) storage account must exist in the same tenant, and the target workspace must be connected to it. For optimal performance, this storage should be in the same region as the workspace capacity.
- Large semantic model storage format must be enabled for the workspace.
- The XMLA endpoint must be enabled for read-write on the capacity. This can be configured in the Fabric Admin portal under capacity settings.
- If on-premises data sources are involved, a Power BI on-premises data gateway must be installed and configured to replicate the AAS gateway setup.
Step-by-Step: How to Migrate Azure Analysis Services to Power BI
The migration process consists of two logical phases: establishing a pairing connection between the AAS server and the Power BI workspace, then executing the model migration itself. Microsoft's tooling handles the heavy lifting — backing up the model, transferring it via ADLS Gen 2, and restoring it into the target workspace — but each step requires correct configuration to proceed.
Phase 1 — Create the AAS-to-Workspace Pairing
Phase 2 — Execute the Model Migration
Server Redirection and Report Rebinding Explained
Two post-migration steps are essential for a fully operational cutover: enabling server redirection and rebinding reports. Both are deliberate — Microsoft does not enable them automatically during migration, giving teams the opportunity to validate the migrated model before switching live traffic.
Server redirection allows applications and tools that are currently pointed at your AAS server connection string to be automatically redirected to the Power BI semantic model. To enable it, navigate to the migration pair under All migrations and toggle the Server redirection enabled slider to Enable. Once active, connections using the AAS endpoint will be transparently routed to the Power BI model without requiring client-side connection string changes — a significant operational advantage for large report estates.
Report rebinding reconnects Power BI reports directly to the migrated semantic model rather than routing through AAS server redirection. To rebind, navigate to Migration Details, locate the newly migrated model, and select Rebind reports. Rebinding is the cleaner long-term state — it removes the dependency on the AAS connection string entirely. Microsoft recommends thoroughly testing the migrated model in Power BI before rebinding reports at scale, particularly for models with complex RLS or custom connection logic.
AAS vs Power BI Premium: Platform Comparison for Leaders
For executives weighing the migration decision against remaining on Azure Analysis Services, the platform comparison below summarises the key operational and capability differences as of 2024.
| Dimension | Azure Analysis Services | Power BI Premium (Fabric) |
|---|---|---|
| Licensing model | Standalone Azure resource — billed separately | Included in Fabric / PBI Premium capacity |
| Microsoft investment trajectory | Maintenance mode — no major new features | Active development — new capabilities quarterly |
| Direct Lake mode | Not supported | Supported — eliminates import latency |
| OneLake integration | Not supported | Native — models read directly from OneLake |
| Git / TMDL source control | Not supported natively | Supported via Fabric Git integration |
| Deployment pipelines | Manual or scripted via XMLA | Native Fabric deployment pipelines |
| XMLA read-write | Supported | Supported (requires read-write capacity setting) |
| Power BI report compatibility | Requires-server redirection or manual-reconnection | Native — no redirection layer required post-migration |
| Backup & restore | Azure Storage — manual configuration | ADLS Gen 2 — integrated with Fabric workspace |
- Microsoft's migration tooling allows you to >migrate Azure Analysis Services to Power BI without rebuilding existing reports - downstream report connections are preserved through server redirection and model rebinding.
- The AAS-to-Power BI migration process uses ADLS Gen 2 as the transfer medium. Workspace and storage must be in the same tenant and, ideally, the same region for optimal performance.
- A maximum of five models migrate concurrently - plan migration batches accordingly for large AAS estates to manage capacity impact on the target workspace.
- Server redirection and report rebinding are separate, deliberate steps — enabling them only after validating the migrated model protects production report quality.
- Power BI Premium on Fabric now offers Direct Lake, Git integration, and deployment pipeline capabilities that are unavailable on AAS — making early migration a capability investment, not just an infrastructure exercise.
Next Steps: Planning Your AAS Migration at Scale
A single-model migration is straightforward to manage internally. An enterprise migration — involving dozens of AAS model databases, multiple workspaces, complex RLS configurations, on-premises data gateways, and hundreds of downstream reports — is a programme that requires structured planning, phased execution, and rigorous post-migration testing before any cutover.
The sequencing decisions alone are consequential: which models migrate first, how gateway configuration is replicated, whether Direct Lake adoption is included in scope, and how report owners are notified and trained to work with the rebinding process all need to be resolved before the first model moves. Getting these decisions right the first time avoids the scenario where a partially migrated estate creates a period of dual-platform complexity that extends the transition timeline and adds risk.
Numlytics has executed AAS to Power BI migrations for enterprise clients across financial services, retail, and professional services — in the US, UK, Australia, and UAE. Our Microsoft Fabric migration practice covers the full journey from migration readiness assessment through phased model cutover, gateway reconfiguration, and post-migration governance using our Power BI Governance Platform. For organisations operating a large AAS estate, we also provide semantic model consulting to evaluate whether Direct Lake adoption should be incorporated into the migration scope — often the right decision when the model is being touched anyway.
If your organisation is evaluating or planning an AAS migration, speak with a certified Numlytics consultant for a no-obligation migration readiness assessment. We will review your current AAS environment, identify prerequisites gaps, and provide a practical migration roadmap tailored to your platform, licensing, and timeline.