Product · SaaS
BizOZ
Bulk document ingestion, made legible while it is still running.
Context
A pile of documents, one searchable record
BizOZ takes business documents — financial reports, conversation transcripts, contracts — and turns them into a typed, searchable record. Analysts upload in bulk, then ask questions across everything they have ingested.
The problem
The middle of the process is invisible
Upload is easy. The hard part is what comes after: files land, get classified, get parsed — and for a stretch the user has no idea what state anything is in. Most tools answer that with a spinner.
An analyst needs three things during that gap: what was recognised, what type it was given, and whether they can start working yet.
Approach
Four decisions
- One canvas. The drop zone, the context field and the record table live on the same screen, so you never lose sight of what is already ingested while adding more.
- Type as a persistent tag. Document type is a colour-coded chip set at upload, carried into the table row, and filterable later. The classification is never hidden inside a detail view.
- State written in words. “Document ingestion transformed into a record successfully” sits next to the progress bar. Language does the reassuring; the bar just does the timing.
- History as a first-class column. Research is iterative, so recent searches and favourites get a permanent sidebar rather than a buried menu.
Visual language
Calm surface for dense work
A soft pink-to-lilac gradient field behind white cards keeps the workspace from feeling like a database admin panel, without adding noise. Rounded 16px cards, wide gutters, and a single violet reserved for primary actions.
