Some features change what you can do. A smaller, better class of feature changes a habit so ingrained you'd stopped noticing it was a choice. OneLake shortcuts are the second kind, and I want to give you three concrete ways they've changed how my team actually delivers reports — not the marketing version, the "here's what we stopped doing" version.

First, the plain translation. A shortcut in OneLake is a pointer to data that lives somewhere else — another workspace, another domain, even another cloud — that lets you use that data in place, as if it were local, without copying it. That's the whole idea. And it attacks the oldest, most invisible waste in data work: the reflex to copy data to wherever you happen to need it.

One: reuse, instead of "just make another copy"

The habit shortcuts broke first was the copy reflex. For as long as I've built reports, the default move when you needed data that lived in another team's world was to copy it into yours — a pipeline to pull it over, a refresh to keep it current, a second physical copy to store and pay for and keep in sync. You did it without thinking, because that's how it worked.

With a shortcut, you point at the data where it already lives and use it. No copy, no sync pipeline, no second refresh to fail at 6am. The first time we deleted an entire copy-and-refresh pipeline and replaced it with a shortcut, it felt almost illicit — like we'd forgotten a step. We hadn't. The step was the waste, and it was gone. Multiply that across an estate and the reduction in redundant pipelines, storage, and things-that-can-break is substantial. Reuse stopped being an aspiration and became the path of least resistance.

Two: trust, because there's finally one copy

The second change was subtler and mattered more. Every copy of data is a chance for two versions of "the truth" to drift apart — the source updates, the copy lags, and now two reports disagree and everyone wastes a morning working out which is right. I've lost genuine days of my life to "why don't these two numbers match," and the answer was almost always "because they're copies that fell out of sync."

Shortcuts collapse that. When five reports all shortcut to the same source instead of each holding their own copy, there is exactly one version of that data, and everything reads from it. The drift problem doesn't get managed better — it stops existing, because the thing that caused it (multiple copies) is gone. That's a governance win disguised as a convenience feature: a single source of truth enforced by architecture rather than by everyone's vigilance and good intentions.

Three: speed, because access stops needing a project

The third change was to how fast we can start. Previously, "we need the sales domain's data in our report" was a small project: request access, build the pipeline, schedule the refresh, test it. Days, sometimes weeks, before a single report got built. With shortcuts and the right permissions, referencing another domain's governed data is close to immediate — you point, you use, you build.

That compression changes what's worth attempting. When accessing cross-domain data was a project, small questions didn't get answered because the setup cost more than the answer was worth. When it's a shortcut, the small questions become answerable, and a surprising amount of value lives in the small questions nobody used to bother with because the plumbing was too expensive.

The best infrastructure features don't give you a new capability. They quietly delete a chore so old you'd stopped seeing it as one — and copying data everywhere was the chore we'd all stopped seeing.

The one I didn't expect: reaching outside Azure entirely

The three above I went looking for. This one surprised me. Shortcuts don't only reference data in other Fabric workspaces — they can point at data sitting in other clouds, in storage outside the Microsoft world, and let you use it in place without copying it into Azure first. The first time we built a report on data that physically lived elsewhere, through a shortcut, with no ingestion pipeline hauling it across, it reframed what "our data estate" even means.

For anyone with a genuinely multi-cloud reality — and that's more organisations than like to admit it — this quietly matters. It means Fabric doesn't demand you relocate everything into its world before you can use it. You can meet your data where it already lives. That's a more honest posture than most platforms offer, and it made me trust the OneLake model more, not less: openness you can actually use beats a walled garden with a nice lawn.

The catch, because there's always one

I'd be writing a brochure rather than a field note if I didn't flag the discipline shortcuts demand in return. Making data trivially referenceable across your whole estate is powerful and it means access and governance matter more, not less. A shortcut is a doorway, and doorways need to be governed — who can point at what, and whether the sensitivity of the source travels with the reference. The convenience is real, but it's convenience that assumes you've done the governance work underneath. Shortcuts reward a well-governed estate and quietly amplify the risks of an ungoverned one, because now the ungoverned data is reachable from everywhere with a click.

What actually changed

Put the three together and the honest summary is that shortcuts didn't add a feature to our reporting — they removed a set of habits we didn't realise were costing us. The copy we always made. The drift we always chased. The access project we always ran. Each of those was a tax we'd paid so long it felt like part of the work, and shortcuts let us stop paying it. That's why they made this list and flashier things didn't: the features that change your daily reality most are usually the ones that let you stop doing something, not the ones that give you something new to do. Use them, govern them, and enjoy the quiet of the pipelines that no longer wake you up.