Where Databricks Lakeflow Designer Ends and Customer File Preparation Begins

Michael Zittermann
Michael Zittermann
Co-Founder & CEO
Last updated on
November 16, 2026

Key takeaways

Lakeflow Designer is Databricks’ visual no-code data preparation tool. Analysts drag and drop steps, or describe what they need in simple instructions, to build a pipeline on data that already lives in Databricks. The goal is to let more people build useful pipelines without overloading data engineers.

It’s a good answer to a real problem inside a company that runs on Databricks. For implementation, onboarding, and operations teams receiving customer files, the first question is where the work starts.

What it is built for

Lakeflow Designer offers a visual canvas with a natural-language assistant:

AspectWhat Lakeflow Designer provides
Who buildsAnalysts, with data engineers able to review and take over the code behind each workflow
OversightGovernance, lineage and permissions from the Databricks catalog
RunningPipelines that can be scheduled, monitored and alerted on
Underlying codeCode behind the canvas that engineers can open, review and edit

For a company that already runs on Databricks, that combination of ease and control is the main appeal. Analysts build, engineers keep the standards, and the data stays inside the governed environment.

It starts inside Databricks

That last point is also the limit for this use case. The design keeps data inside the platform. A customer’s file is outside it.

To be fair, Designer does take in files:

  • You can drag in CSV, JSON, Excel or PDF files, or pull them from Google Drive or SharePoint.
  • A validation step separates rows that pass its checks from rows that fail.

However, the file has to land in Databricks first. A local file is uploaded to a Databricks storage area, so the person handling it needs a workspace and the right permissions. Databricks doesn’t appear to position Designer for an implementation team receiving files from many customers, each with a different format.

What the setup assumes

Designer assumes a Databricks workspace and an analyst who works in it. Most implementation teams at software companies don’t. Their daily tools are the customer’s spreadsheet, a template, and the system they are loading data into.

Asking an implementation team to learn a lakehouse architecture just to handle customer files creates unnecessary overhead. Their core work happens much closer to the customer than to the data platform.

Who owns pipeline changes

In Lakeflow Designer, data engineers can inspect and edit the underlying code, while analysts build their own workflows—creating a clear division of labor for internal data projects.

However, customer file ingestion rarely follows a predictable internal schedule. When a customer changes a file layout or data structure, someone must adapt the pipeline immediately.

Because account managers and implementation teams usually lack Databricks workspace access, even minor format updates require submitting tickets, waiting on data engineering cycles, and explaining delays to the customer.

How data intake works in Ingestro

Ingestro lets implementation teams build and manage data flows visually, without relying on data engineers to develop custom pipelines.

When a new file arrives, AI suggests how its columns map to your target data model based on meaning, not just exact header matches. Teams can review the suggestions in a spreadsheet-like grid, fix an entire column with a simple plain-language instruction, and validate the data against their own systems.

You build the workflow once on a visual canvas. From there, every new file runs through the same process, either on a set schedule or as soon as it arrives.

Ingestro handles multi-tab Excel workbooks, PDFs, and XML files, and automatically adapts to schema changes. You can also embed the importer directly into your web application to give customers a native import experience.

With the embedded importer, data preparation takes place in the browser, keeping sensitive data off external servers during preprocessing.

Ingestro offers US and EU cloud hosting as well as self-hosted deployments. It’s ISO 27001 certified, and Business plans include dedicated solutions engineering support.

Where the preparation time goes

Designer provides general steps for joining, pivoting, and renaming, letting you preview adjustments before applying them. Those steps suit a data team that knows its table structures in advance. A customer file is different: it often arrives with nested data, swapped rows and columns, or content spread across several files, mapping to hundreds of potential options.

Validation and cleaning must then follow, sometimes against a separate database. That workload isn’t easy to manage through a chat interface or general-purpose canvas. Ingestro was designed for this exact sequence, with dedicated screens for preparation, mapping, and review so teams can handle the entire file in one place.

Where the two fit together

If your company already runs on Databricks, the two tools can complement each other. Ingestro cleans and standardizes customer files first, allowing the structured data to flow into the lakehouse where Designer and analysts can take over.

Checks to run before you commit

  • Where does the work begin: with files from outside customers, or with data already in Databricks?
  • Do the people who handle customer files have a Databricks workspace, or the time to learn one?
  • When a customer changes a file, can the account owner fix the workflow today?

Choose the customer file your team struggled with most this quarter, and ask each vendor to show the path from that file to clean data.

Frequently asked questions

Ingestro report

The hidden costs of implementation
Nearly 200 software companies across 12 industries reveal where implementation and onboarding teams lose the most time.
Read our report

See how easily you can turn your customer files into ready-to-use data with Ingestro’s AI agents.

Keep exploring

icon