Product · AI & automation
How the AI Conversion Works
What the AI sees in your document, how it decides what becomes a field, and how to get the best results.
When you upload a document, the AI reads it the way a careful human would: labels, blanks, checkboxes, tables, signature lines, and the relationships between them. It then rebuilds that structure as native form fields with sensible types. This section explains what is happening under the hood, what kinds of documents convert best, and how to phrase and prepare documents for the cleanest possible output.
Who this is for
- Users curious why the conversion is as good (or occasionally odd) as it is
- Teams preparing batches of documents for conversion
- Skeptics who want to know where AI helps and where humans stay in charge
Key things to know
Layout is language
The AI infers intent from layout: a label beside a blank line means a text field, a row of boxes means options.
Cleaner input, cleaner output
Straight scans, legible text, and consistent formatting convert with fewer corrections.
You have the final say
AI proposes; you approve. Nothing publishes without your review.
Quick answers
What does the AI do with handwriting on a scanned form?
It reads the printed structure of the form. Existing handwritten answers are ignored, since the goal is a blank fillable form, not a transcription of one filled copy.
Will the AI change the wording of my form?
No. Labels and questions are preserved as written. The AI translates structure (blanks, boxes, tables) into field types, not your language into its own.