How it works

A predictable path from files to output

Each step is explicit and reversible. Your original data stays untouched from the first upload to the final export.

1
Create project
A project is the container for one body of work: its files, mappings, review decisions, runs and exports. Retention is set here.
2
Upload source files
Drop in a master list plus any related files. Uploads are stored as immutable source objects; nothing is parsed destructively.
3
Identify file purpose
Tell Informatt what each file represents. Purpose drives default match keys and how returned results are folded back in.
4
Inspect columns
Each column is profiled — inferred type, fill rate, sample values — so you can spot problems before they propagate.
5
Map fields
Map source columns to a target schema. Suggestions are ranked by confidence and always require confirmation.
6
Match & reconcile
Records are joined across files using the keys you choose. Deterministic matches resolve first; the rest are scored.
7
Review ambiguities
Anything below your confidence threshold enters a review queue with both candidates shown side by side.
8
Choose transformation
Apply normalization and formatting rules as a saved recipe: casing, phone regions, postal formats, derived fields, row filters.
9
Preview output
Inspect the exact output — headers, order, formatting, row count — before generating a file.
10
Export
Generate the file in the layout the receiving system expects. Re-run the same recipe whenever new data arrives.

Walk through it with sample data

The workspace is populated with example projects so you can see every screen before uploading anything.