Your platform team tells a great story: all of your data in one lakehouse, one governance model, ready for analytics and AI. It's a story they've spent years and millions making true.
Meanwhile, back at the office, hundreds of your analysts will do what they do every morning: pull that same data out of Databricks and into Alteryx workflows on laptops and a Server cluster down the hall.
Your data has a commute. Every day it leaves a governed, infinitely scalable platform to go work on a desktop — and everything you built the lakehouse for stays behind. The problem isn't your analysts, and it isn't the lakehouse. It's the desktop-era tool shuttling between them. Here's what the commute costs.
The lakehouse promise, and the business teams locked out of it
Be fair to your analysts: they didn't choose this commute to spite the architecture. The lakehouse speaks SQL, PySpark, and notebooks. Most business users don't. When the front door requires code, people find a side door, and for a decade, Alteryx has been the side door of choice: drag, drop, done.
The result is an irony most enterprises live with quietly: the most data-hungry people in the company do their work outside the platform that was built for them. Finance, ops, and risk teams — the heaviest consumers of governed data — are the least connected to its governance.
The lakehouse promise stops at the business user's desk.
If the data stayed home: Prophecy gives business users a visual canvas on Databricks itself, and the side door becomes a front door.
Commute cost #1: Governance ends at the export
Unity Catalog knows who can see every table. It tracks lineage, tags PII, and writes the audit trail your compliance team depends on. All of that holds right up to the moment an analyst clicks “run” and the data lands in a workflow file on a desktop.
From there, nobody knows. Which laptop holds customer data? Which Alteryx output fed which report? Who emailed which extract to whom? Every export is a copy your governance can't see — and at enterprise scale, it's thousands of copies a week.
This is the paragraph your next security review writes for you. Better to read it here first.
If the data stayed home: Prophecy workflows execute inside Databricks under Unity Catalog. The data never leaves, so governance never ends.
Commute cost #2: Multiple sources of truth
Two analysts extract the same revenue table — one on Monday, one on Thursday. Each applies slightly different prep logic. Two numbers walk into next week's exec meeting, and the meeting becomes an investigation: thirty minutes spent reconciling instead of deciding.
A single source of truth requires that the business logic and the data live in one governed place. Desktop workflows fork both: the data forks at export, and the logic forks across hundreds of .yxmd files that no one can review or reconcile.
The damage compounds quietly. Once numbers stop matching, every dashboard gets re-litigated — and trust, once spent, is expensive to buy back.
If the data stayed home: with Prophecy, visual prep logic lives as versioned open code on Git, in one place.
Commute cost #3: The scale wall
A desktop engine can process roughly what fits on a desktop. You know your teams have hit the wall when you hear the workarounds: "we sample it first," "we split it into four workflows," "we kick it off Friday and check Monday."
Here's the absurd part. While that laptop churns overnight, Photon and elastic compute sit idle in the platform you're already paying for — engineered for exactly this job, locked out of it by the tool's architecture.
AI-era data volumes only move the wall closer. The fix isn't a bigger desktop, but rather, putting an end to the commute.
If the data stayed home: Prophecy runs execution natively inside Databricks — your workflow scales because the lakehouse does.
Commute cost #4: Paying to run a second platform
You already operate a production data platform: Databricks. Scheduling, monitoring, scaling, security, and support are solved and budgeted.
Alteryx Server is a second one. Separate licenses, separate infrastructure, separate admins, separate patch cycles, separate support contract — commonly a six-figure line item before anyone builds anything. Two production systems mean double the operational surface, and one of them exists mostly to run logic the other could run natively.
Renewal season makes the question concrete: what exactly is Server doing that the platform you already pay for can't?
If the data stayed home: Prophecy runs and schedules on the platform you already operate — no second production system to license or babysit.
Commute cost #5: The rewrite tax
Trace the life of a workflow that matters. An analyst builds it in Alteryx. It proves valuable, so it needs to run reliably, at scale, in production, which means data engineering rebuilds it in PySpark or dbt on Databricks. Now there are two versions of the same business logic, written by different people in different languages, drifting apart.
You pay for the logic twice and then maintain both forever. Within a quarter, the analyst's version and the production version disagree, and nobody is sure which one is right.
Here's the reframe that should sting: you're already migrating off Alteryx. Continuously, one expensive rewrite at a time, in the wrong direction.
If the data stayed home: in Prophecy, the analyst's workflow is already production code with tests, versioning, and CI/CD.
What it looks like when the data stays home
Business users build visually — the ease that made Alteryx win, on a canvas that lives where the data lives. Under the hood, every workflow is open code on Git: versioned, reviewable, testable, one source of truth for logic and data alike. Execution pushes down to Databricks, under Unity Catalog, at whatever scale the lakehouse offers that day. And because the visual workflow is production code, there is no second platform to operate and no rewrite tax to pay.
AI raises the ceiling further: it drafts workflows and reads messy spreadsheets in seconds, while humans inspect, refine, and validate visually. AI delivers speed, while maintaining trust.
That's Prophecy: AI-native data preparation with low-code trust, running inside your lakehouse.
Stop renewing the commute
This was never really a tooling decision. You made the architecture decision when you chose the lakehouse; Alteryx is the last desktop-era holdout still commuting against it.
Want to learn more? Check out our ebook on Building Data Pipelines on Databricks in 5 Easy Steps.
See Prophecy in Action
Request a demo to see how Prophecy eliminates the data commute and helps your teams get the most out of your Databricks investment.
