If you work with data from vendors, partners, or internal systems, you know the drill: a file arrives in one format, and you need it in another. Column names don't match. Dates are in the wrong format. Fields need splitting, merging, or filtering.
Most teams solve this one of three ways — all of them painful.
The three bad options
Option 1: Do it manually in Excel. Open the file. Rename columns. Write formulas. Copy-paste. Every. Single. Time. It works until you're doing it three times a week and losing your mind.
Option 2: Write a Python script. This works great — until the vendor changes their format, the script breaks at 2am, and nobody remembers how the script works. Now you're debugging someone else's regex at midnight.
Option 3: Buy an enterprise ETL tool. Talend, Informatica, or Azure Data Factory can absolutely do this. They can also take months to set up, cost tens of thousands per year, and require dedicated engineers to maintain.
None of these are good answers for the 80% of data transformations that are straightforward but repetitive.
What if you could just describe what you want?
That's the idea behind Data Shepherd. Instead of writing code or dragging boxes around a canvas, you describe the transformation in plain English:
"Combine first_name and last_name into full_name, extract the domain from the email column, and reformat date from MM/DD/YYYY to YYYY-MM-DD."
The AI generates a deterministic transformation script, shows you a preview of the results, and lets you approve before running on your full dataset.
No Python. No formulas. No IT tickets.
How it actually works
-
Upload your file — CSV, JSON, TXT, or fixed-width. Data Shepherd analyzes the structure automatically.
-
Describe your transformation — Use the AI agent chat to explain what you need. You can refine iteratively: "Actually, also filter out rows where status is inactive."
-
Preview the results — See exactly what the output looks like before committing. If something's off, adjust your description.
-
Save and reuse — The transformation is saved. Next time you get the same file format, run it again in one click. If the source format changes slightly, auto-heal adjusts the script automatically.
Who is this for?
Data Shepherd is built for the people stuck in the gap between "I can use Excel" and "I need a data engineer":
- Operations managers who receive vendor files weekly and spend hours reformatting them
- Data analysts who know what they want but don't want to maintain Python scripts
- Small teams that can't justify an enterprise ETL platform for straightforward transformations
- Developers who are tired of writing one-off scripts for non-technical teammates
The bottom line
If you're spending more than 30 minutes a week manually transforming data files, there's a better way. Describe what you need, let AI handle the transformation, and get back to work that actually matters.
Try Data Shepherd free — no credit card required.