The Open-Source Case for Data Prep

Here's something about Alteryx that does not show up on your invoice: after enough investment in their product, you are a full-fledged hostage to them. Just ask Amgen's finance team. They built 200+ workflows the business required, and every single one of them lived in a proprietary desktop app format that would only run in the software they were trying to leave.

Amgen teamed up with Prophecy and Databricks and rebuilt those 200+ Alteryx workflows as open code, on the cloud platform their team already owned. Those workflows now feed 500 dashboards and more than 1,500 people inside the company, all running on infrastructure the platform team trusts.

That kind of clean exit is rare. Most data prep tools are engineered so you can't do what Amgen did: proprietary runtimes, proprietary file formats, per-seat pricing that climbs with every new analyst. Prophecy was built to flip that arrangement on its head. Give analysts the visual, drag-and-drop experience they want, but have it compile down to open code that runs on the data platform you already own. And as we've learned building agentic data preparation, that story tends to land with everyone in the room for different reasons.

What open source means at Prophecy

At Prophecy, open is more than a marketing term. Every visual workflow compiles to standard, open code (SQL or Spark) and commits right into your Git repository.

Concretely:

• Every visual workflow is code. Analysts drag and drop; Prophecy generates open code any engineer on your team can use.
• The code lives in your Git repository. Not in a vendor's metadata store. Your engineers can review, test, version, and run it long after any contract ends.
• Execution happens on your platform. Prophecy deploys that code directly to Databricks, Snowflake, or BigQuery. Compute, security, and governance stay with your platform team.

If you ever leave Prophecy, you keep working. That matches what buyers say they want: in a Forrester Total Economic Impact study of 217 IT, data, and platform decision-makers, 86% rated interoperability across engines and clouds as important or critical.

The Alteryx problem: proprietary workflows, proprietary economics

Alteryx pioneered visual data prep, and analysts like the interface. That affection is what makes lock-in painful when the bill arrives or the platform strategy shifts.

Independent reviewers keep landing on the same issues:

• Migration is expensive. A comparison analysis notes that Alteryx's proprietary format makes it hard to move workflows to other platforms or adopt new tools without significant rework.
• The GUI is opaque. Alteryx's visual layer hides what's happening under the hood, slowing code review and debugging compared to transparent, code-based alternatives.
• Per-seat pricing scales badly. G2 lists Alteryx starting at $3,000 per user per year for its Starter Edition, with enterprise editions rising sharply from there.

The commercial picture shifted again in March 2024, when Clearlake Capital Group and Insight Partners acquired Alteryx for $4.4 billion. The transaction closed in March 2024, taking the company private and adding fresh uncertainty for regulated buyers around versioning and cloud data-processing changes.

Prophecy inverts the model. Because every workflow is open code from the first save, there is nothing to reverse engineer if you leave. Analysts get the same visual, low-code experience they know from Alteryx, without paying for it in optionality. For teams already planning a move, our Alteryx migration playbook walks through the mechanics.

The Informatica problem: opaque pricing, consolidation risk

Informatica sells freedom of choice, while customer reality is different.

A 2026 Informatica pricing guide that pulled together Reddit threads, G2 reviews, and industry forums summarizes the pattern plainly: Informatica publishes no standard rates or tiers, and even experienced data professionals struggle to get straight pricing during sales calls. The same guide walks through real customer totals where a starting PowerCenter deal in the mid-six figures balloons past $600K by year three once maintenance and additional connectors are added.

Then there's the consolidation risk. Reacting to Salesforce's $8B acquisition of Informatica, the Orchestra Data Leadership Newsletter framed it as vendor consolidation and lock-in reaching a new peak. The more you build on Informatica, the more your data strategy depends on Salesforce's roadmap.

Why the open-source approach works for everyone

Each persona values a different consequence of the same open architecture.

Analysts think: the tool I rely on won't get taken away from me. The workflow they build today survives the next platform migration because the underlying code is portable.

Platform teams think: the tool won't create a shadow stack. Prophecy deploys to Databricks, Snowflake, or BigQuery with governance, version control, and lineage that stay in the platform team's hands.

Analytics leaders think: the switching-cost bomb is defused. The line item for Alteryx or Informatica covers more than this year's license. It carries the accumulated liability of every workflow trapped in a format only that vendor can execute.

Open code is only half the story: governance is what makes it work

Portable code without governance just moves the problem. Analysts still need to ship trustworthy work, and platform teams still need to keep the environment clean. Prophecy pairs open output with governed self-service: analysts move fast on their own, inside guardrails the platform team defines once, with governance baked into the artifact from the start.

Here's what that unlocks for each persona:

This is how Prophecy brings governed self-service to agentic data preparation. Analysts build visual data workflows on the cloud data platform, AI agents draft a first-pass workflow that a human refines and validates before production, and every workflow compiles to open code your team can read, own, and take with them.