The AuditBuilder (Light): Overview and FAQ
The AuditBuilder (Light) helps you set up a data analysis quickly and easily — no technical knowledge required. Using AI and carefully designed analysis templates, you can generate both a script and the corresponding audit documentation in just a few steps.
The current Light version is intentionally simple and accessible. You select an analysis, map your data fields, and set a few parameters. You’ll then receive a fully reproducible Python script and a ready-to-use audit record. If users find this helpful, a full version with extended functionality may follow in the future.
Frequently Asked Questions
What do I need to get started?
You only need the datasets listed in the selected template (for example, 'invoices' or 'bookings') and the column names as they appear in your files. No data upload is required — simply enter your column names and parameters, such as an account number or threshold value.
Which file formats are supported?
For each source file, you can choose between XLSX, CSV, or JSON. The AuditBuilder automatically adjusts the import section of the generated script to match your choice.
How does the field mapping work?
Each template includes example fields. For every field, you enter your own column name. The generated script will rename or select these columns so the analysis can run directly on your dataset.
Can I edit or extend the results later?
Yes. The AuditBuilder creates a complete Python script and a Markdown documentation file. You can edit, extend, and rerun both at any time. Use the follow-up prompts on the page to refine the script further if needed.
How is my data kept private?
You don’t need to share any data. The prompt only contains structure information — column names, parameters, and minimal context. All processing happens locally, and you retain full control over your data.
I’m getting an error when generating the output. What should I do?
Check whether all required columns are filled in and whether the parameters are of the correct type. Review any warnings above the script, or ask a follow-up question with your error message for assistance.