> ## Documentation Index > Fetch the complete documentation index at: https://docs.abbyy.com/llms.txt > Use this file to discover all available pages before exploring further. # Extract data from documents in Advanced Designer > Choose the right Advanced Designer scenario based on document structure — structured, semi-structured, unstructured, or mixed. To extract data from documents in Advanced Designer, create a [Document skill](/vantage/documentation/advanced-designer/document-skills/document-skills). The activities you use depend on the structure of your documents — Vantage groups documents into four broad categories. ## Document types ### Structured documents **Structured documents** (also known as fixed forms) always carry the same fields in the same layout, or in a very small number of layouts. Examples: forms, questionnaires, surveys. Sample structured document — a form with consistent field positions ### Semi-structured documents **Semi-structured documents** carry similar information across documents, but field locations, sizes, and counts vary. Vantage uses spatial and logical relationships between elements to locate the data. Examples: invoices, payment orders, bills of lading. Sample semi-structured document — an invoice with varying field positions ### Unstructured documents **Unstructured documents** are freeform text in paragraphs and sentences, where the fields you need may span pages. Examples: contracts, emails, research articles. Sample unstructured document — paragraphs of freeform text ### Mixed document sets A **mixed document set** contains documents of different types — for example, semi-structured and unstructured together, or single documents that mix the two (paragraphs of plain text alternating with tables). ## Pick a scenario Once you've identified your document type, follow the matching scenario. Build skills for forms with fixed or limited layouts. Build skills for invoices, bills of lading, and similar varying-layout documents. Build skills for contracts, emails, and other freeform text. Handle document sets that combine multiple structure types.