Upload & parse

The first job is to get the questionnaire into SolveGRC and turn it into a list of individual questions the AI can answer. You upload the file, then parse it. Parsing reads the document and extracts each question into its own reviewable row.

Upload the questionnaire

From the Questionnaires hub, click New questionnaire. (The button only shows if your role can create questionnaires.)

The New-questionnaire dialog: a name field, an optional description, and a file picker showing the accepted formats.

In the dialog:

1. Give the questionnaire a name you will recognize in the list, usually the customer or framework, like "Acme Corp, Vendor Security Review."
2. Add an optional description for context.
3. Choose the file. SolveGRC accepts Excel (.xlsx), CSV (.csv), PDF (.pdf), and Word (.docx) files up to 25 MB.
4. Click Upload.

The questionnaire lands in the list as Draft: uploaded, but not yet read.

Use the spreadsheet version when you have it

Most security questionnaires (CAIQ, SIG, and vendor templates) arrive as spreadsheets, and spreadsheets are the format SolveGRC reads most reliably. If a customer sends a PDF but a spreadsheet version exists, ask for the spreadsheet. You will get cleaner question extraction with no OCR step in between.

How Word documents are read

.docx questionnaires parse without an OCR step. Tables in the document are read with their structure intact — each table gets the same header and column detection as an Excel sheet — so table-based questionnaires (the common Word layout) extract cleanly. A Word document with no tables is read as prose, with questions detected line by line like a text PDF.

Parse it into questions

On the Draft card, click Parse. For a spreadsheet, extraction runs right away: SolveGRC opens the workbook, works out which rows are questions, and pulls each one into the questions table, numbered, typed (yes/no, multiple choice, free text), and with any answer options it finds. Duplicate questions are collapsed automatically.

While it works, the card shows a live progress indicator with the current stage and, for multi-sheet workbooks, a per-sheet bar.

Parsing complete: the card flips to Ready with its extracted question count — here 12 questions read from an uploaded CSV, with a PDF still waiting as Draft below.

When extraction finishes, the status flips to Ready, and the questions are in and waiting for answers.

PDF questionnaires

A PDF has no spreadsheet structure to read, so parsing one first asks how you want its text extracted. Fast does a quick text extraction and suits clean, digital PDFs. Advanced (Forms & Tables) runs a deeper pass that understands form fields and tables, for scanned or form-heavy documents.

The OCR method chooser shown when you parse a PDF: Fast versus Advanced (Forms & Tables), each with its expected processing time.

PDF extraction is best-effort

Reading questions out of a PDF is inherently less precise than reading a spreadsheet. Give the extracted questions a careful look on the Ready questionnaire, and prefer a spreadsheet source whenever one is available.

Check the extracted questions

Click a questionnaire card to open its detail view. You will see the status timeline, the sheets that were read, and the full list of extracted questions. This is a good moment to confirm the extraction caught everything and did not pull in stray rows from an instructions or cover sheet.

The questionnaire detail dialog: file information, answer progress, the per-sheet breakdown, and the status timeline.

If parsing fails

If SolveGRC can't read the file, the questionnaire returns to Draft and a message explains why. Nothing is lost. Fix the issue (wrong format, a corrupt or password-protected file, a scan with no readable text) and click Parse again.

---

Once a questionnaire is Ready, you can draft answers. Continue to Generate answers.