Fabric and Azure

Legacy BI to Microsoft Fabric: A Migration Guide

Moving legacy BI to Microsoft Fabric without breaking trusted reports. A plain guide for a Head of Data on what to move first and how to plan it.

Austin Levine

·

Updated

Moving legacy BI to Microsoft Fabric means retiring old reporting tools and rebuilding your reports on one governed platform where every number has a single definition and a clear owner. The goal is to fix the model underneath so both people and Copilot return the same trusted number, instead of copying your old dashboards straight across.

Your old reporting tools still run the business, and every plan to move legacy BI to Microsoft Fabric slips to next quarter. The reports mostly work, nobody wants to break them, and the person who understood how they were built is either very busy or already gone. So the migration stays on the roadmap and never leaves it.

This is a plain guide to what the move actually involves and how to do it without the numbers breaking in front of the people who rely on them. If you're a Head of Data being asked when the old platform gets switched off, read this before you commit to a date.

Key takeaways

  • Moving legacy BI to Fabric retires old tools and rebuilds reports where every number has one definition and owner.

  • Fix the model underneath so people and Copilot return the same number, instead of copying old dashboards across.

  • Start with an inventory, not a rebuild, so you see what you actually have before moving it.

  • Move the highest-risk reports with the clearest definitions first, like the monthly finance pack.

What does moving legacy BI to Microsoft Fabric actually mean?

Legacy BI is the reporting setup you already have: an older tool such as a QlikView site, a Cognos server, a Tableau estate, a SQL Server Reporting Services box, or a pile of scheduled Excel exports that feed the monthly pack. Microsoft Fabric is one platform that keeps your data, your models, and your reports in a single place called OneLake, with Power BI as the reporting layer on top.

So the move is more than swapping one tool for another. In most legacy setups, the logic that decides what a number means lives inside individual reports. One report calculates revenue one way, another does it slightly differently, and there's no single place that says which is right. When you move to Fabric, you pull that logic out of the reports and into one semantic model that every report and every AI answer reads from.

That's the real work. Copying old dashboards across gets you a new tool with the same old problems. Rebuilding the model gets you numbers people can trust.

Why do legacy BI to Microsoft Fabric migrations stall?

Because the risk feels bigger than the reward, and for a while it is. The old reports are wrong in ways the team has learned to correct by hand. Nobody wrote down those corrections. So a straight rebuild produces numbers that look different from the old ones, finance asks why, and the whole project loses trust in week two.

Here's how it usually plays out. A team starts the migration, rebuilds the top ten reports, and shows them side by side with the old ones. Three numbers don't match. The differences are actually the new reports being correct and the old ones carrying an error everyone forgot about, but there's no documentation to prove it. The safe move looks like keeping both systems running until someone can explain the gap, and now you're paying for two platforms and trusting neither.

More caution doesn't fix this. What fixes it's knowing, before you touch anything, which numbers are supposed to match and which old figures were wrong all along. That comes from writing down the definitions first, and it comes before you rebuild.

How do you plan a legacy BI to Microsoft Fabric migration?

Start with an inventory, not a rebuild. You need to see what you actually have before you move any of it.

First, list every report on the old platform and mark which ones people really use. In most estates, a large share of reports haven't been opened in months. You don't migrate those, you retire them. Moving dead reports is how a three-month project becomes a year.

Second, for each report that stays, write down what its key numbers mean. Revenue, active customers, margin: one definition each, agreed with the people who use them. This is the step teams skip and the reason migrations stall, so do it before you build.

Third, pick the pattern for each source. Some data lands in OneLake through a pipeline in Fabric, some connects live, some gets rebuilt from scratch because the old logic was never clean. Naming the pattern per source stops surprises halfway through.

Fourth, move in slices, not all at once. Pick one report area, rebuild it in Fabric on the agreed model, run it in parallel with the old one until the numbers tie out, then switch that slice off on the old platform. Repeat. Running both for a short, planned window per slice is normal. Running both indefinitely because nothing is finished is the mistake to avoid.

Step

What you decide

Why it matters

Inventory reports

Which reports stay, which get retired

Stops you migrating reports nobody opens

Define the numbers

One agreed meaning per key metric

The old reports disagree, so pick the truth first

Choose the source pattern

Pipeline, live connection, or rebuild

Avoids surprises halfway through the move

Migrate in slices

One report area at a time, run in parallel

Keeps trusted numbers live while you switch

What should you move first?

Move the reports that carry the most risk and have the clearest definitions, not the flashiest dashboards. The monthly finance pack, leadership numbers, the figures that go to a bank or an auditor: these are where a wrong number costs the most, so they earn the careful rebuild first.

Leave the exploratory and one-off reports for last, or retire them. A dashboard three people glance at once a quarter doesn't need to be first in line, and some of it shouldn't be rebuilt at all.

Microsoft Fabric passed 28,000 customers by late 2025, the fastest-growing analytics product in Microsoft's history (Microsoft, Ignite 2025). That growth is why the platform is a safe target to build on, but it also means a lot of teams are moving at once and copying old messes straight across. Being deliberate about what you move first is what separates a clean migration from a rushed one.

How long does it take and what does it cost?

It depends on how many reports actually matter and how much of the definition work is already done. A focused migration of the reports that run the business is a matter of weeks to a few months, not a single big switch. The inventory and definition work takes days and saves you far more than that later, because it stops you rebuilding things nobody needs.

The cost that surprises teams isn't the platform, it's running two systems at once for longer than planned. Every extra month of parallel running is a month of paying for the old license, the new capacity, and the people keeping both running. A tight plan with clear slices is the thing that keeps that window short.

If you want to see where your current setup stands before you commit to a date, our free Power BI Governance and Cost Scorecard walks you through the key readiness questions in about five minutes and gives you your top risks. If the score shows gaps, book a migration review with us, and we'll map exactly what to move first, what to retire, and how to keep the numbers trusted while you switch.

FAQ

Can you just copy your old reports straight into Microsoft Fabric? You can, but you'll bring every disagreement and hidden error with you and end up with a new tool that people trust as little as the old one. The value of the move comes from rebuilding on one agreed model, not from a copy. Retire the reports nobody uses and rebuild the ones that matter properly.

Do you have to move everything off the legacy platform at once? No, and you shouldn't. Migrate one report area at a time, run it in parallel with the old version until the numbers tie out, then switch that slice off. The one thing to avoid is running both platforms indefinitely because nothing is ever finished.

What's the hardest part of a legacy BI to Microsoft Fabric migration? Agreeing what the numbers mean. The old reports usually disagree with each other, and the corrections people apply by hand were never written down. Settling one definition per key metric before you build is the step that makes the whole migration reliable.

Read next: Microsoft Fabric: What It's and When to Move and Tableau to Power BI Migration Guide.

Is your Power BI set up the right way?

1

Book 45 minutes and tell us what you want looked at.

2

We walk through your reports and setup with you, live.

3

Within two working days, you get our plan in writing.

Is your Power BI set up the right way?

1

Book 45 minutes and tell us what you want looked at.

2

We walk through your reports and setup with you, live.

3

Within two working days, you get our plan in writing.

Is your Power BI set up the right way?

1

Book 45 minutes and tell us what you want looked at.

2

We walk through your reports and setup with you, live.

3

Within two working days, you get our plan in writing.

CaseWhen is a Berlin BI consultancy that builds reporting that leaders can trust, on the Microsoft stack: Power BI, Fabric and Azure.

Berlin, Germany

© CaseWhen Consulting GmbH

English

CaseWhen is a Berlin BI consultancy that builds reporting that leaders can trust, on the Microsoft stack: Power BI, Fabric and Azure.

Berlin, Germany

© CaseWhen Consulting GmbH

English

CaseWhen is a Berlin BI consultancy that builds reporting that leaders can trust, on the Microsoft stack: Power BI, Fabric and Azure.

Berlin, Germany

© CaseWhen Consulting GmbH

English