Scalefree and IRiS on migrating to Microsoft Fabric without carrying the mess across.
Most organisations moving to Microsoft Fabric are not moving because of the platform. They are moving because of what the platform is supposed to make possible next: AI on their own data, copilots that answer real questions, agents that can act on something trustworthy.
That is the destination. The problem is how most migrations arrive.
The fastest way to get onto Fabric is to do a basic lift and shift. Take the existing warehouse and pipelines, move them across broadly as they are, and call the migration done. It is quick, the upfront cost is low, and when the clock is running it feels like the safe choice.
What it does not do is leave you better off. The tangle travels with you. The data underneath is still hard to find, inconsistent and undocumented, only now it sits on a newer bill. Six months after everyone celebrated the migration, the thing the migration was for has quietly stalled, because AI cannot act on data it cannot understand.
The unpacking was always the slow part
There has always been an honest reason teams choose a basic lift and shift, and it is almost never ignorance. It is time.
Rebuilding the integration layer properly on the way in is the right answer, and everyone knows it. It is also the part that eats the calendar. Designing new models and writing new code for every source, enforcing standards consistently across teams, documenting definitions as you go; that work has historically taken months. Nobody wants to hold a migration open that long, so the trade off gets made, and the boxes get tipped in.
That trade off is what we set out to remove.
What changes when the integration layer builds itself
IRiS is a Lakehouse Automation platform. It reads your source metadata, profiles the data and guides you through the data model design, then generates the Silver layer for you: production ready and operational metadata, with modelling guardrails enforced automatically for you to review and approve. Under the hood, it uses Data Vault 2.1 to structure the integration layer, which gives you stable entity identity, complete history, explicit relationships and an ontology your AI can use.
In practice, that means the unpacking happens at the speed of the move. You still migrate to Fabric quickly. You just arrive with an organised, governed, AI ready foundation instead of a backlog.
IRiS takes the data modelling and code development work that, until very recently, any optimistic scope would put at four hours per table and reduces it to under 15 minutes. 10x faster modelling and coding is what makes a genuine AI ready migration possible.
Resimac, an Australian financial services organisation, put that to the test on a real migration. Using IRiS, they migrated and integrated a newly acquired business into their platform, delivered under time, under budget, and with zero defects.
Because the change is confined to the layer underneath, the reports and dashboards people depend on keep working while it happens. Nobody has to live out of boxes.
Scalefree and IRiS
Scalefree International combines global Data Vault leadership with Microsoft Fabric expertise, driving fast, frictionless platform migrations and providing the end-to-end solutions and training required to guarantee high data quality, empowering organisations with a robust, AI ready platform.
Together, Scalefree and IRiS cover both halves of the problem. Scalefree brings the migration expertise and the practitioner capability. IRiS removes the hand coding that makes doing it properly feel unaffordable.
IRiS is available on the Microsoft Azure Marketplace as a Managed Application and is MACC eligible, so purchases draw down against your existing Microsoft Azure consumption commitment rather than going through a separate procurement process.
Ready to build, or ready to unpack
None of this is an argument about whether to move to Fabric. Nearly everyone is, and nearly everyone should. It is an argument about how you will arrive.
If you are planning a migration, we would rather show you what this looks like on your own data than in the abstract.
Meet us at dataMinds Connect in Mechelen, 12 to 14 October.
.png?)
