Product · Import & document analysis
From PDF, Word, Excel, image, audio, to review-ready requirements
The import pipeline converts files, runs AI analysis against project context, and creates candidate stakeholder requirements with source traces. You stay in control before requirements enter the project base.
Overview
Upload creates an import job linked to the project.
Conversion (fileConverter) produces Markdown or structured ZIP for advanced PDF/DOCX.
Analysis (fileAnalysis) creates requirement areas and candidate stakeholder requirements with text, validation criteria, confidence, and position info.
Review in UI, approve, merge, or adjust before completing the import.
Formats and key capabilities
Multi-format, Office, PDF, image, audio, data
Office: .docx, .doc, .pptx, .ppt, .xlsx, .xls. PDF (including structured PDF→ZIP for pages/tables). Images with OCR: .jpg, .jpeg, .png, .gif, .bmp, .tiff, .webp. Audio with transcript + AI meeting summary: .mp3, .wav, .m4a, .flac, .ogg, .wma, .aac. Data/text: .txt, .md, .html, .htm, .csv, .json, .xml.
AI decomposition to stakeholder requirements
The default flow sends converted content to /api/jobs/intressentkrav with project context, language, and import parameters. Advanced PDF/DOCX can go via intressentkrav-advanced-pdf with a structure ZIP.
Version comparison
On isVersionUpdate version-compare compares previous source against new rows with embeddings and classifies matches (unchanged, modified, new, uncertain, removed). Results are stored on the import file for UI.
Saved import instructions
ImportInstruction lets you save and reuse natural language that steers how AI interprets similar documents, reducing variance across teams and deliveries.