Back to guide/General Productivity

Data Extraction from Documents

Gemini's multimodal document understanding handles PDFs, images of forms, and scanned documents. The explicit normalization rules and NOT_FOUND convention make the output immediately usable in spreadsheets or databases.

gemini-promptsdocument_typefields_listoutput_format
Edit View
Prompt
Extract structured data from the attached {{document_type}}.

Extract these fields:
{{fields_list}}

Output as a {{output_format}} with one row per {{entity_unit}}.

Extraction rules:
- If a field appears multiple times, take the most recent / most specific value
- If a field is missing, use "NOT_FOUND" (not null, not empty)
- For dates, normalize to YYYY-MM-DD format regardless of input format
- For currency, normalize to numbers without symbols (include a "currency" column)
- For names, use "LastName, FirstName" format
- Flag any field where the extraction is ambiguous with a trailing " [AMBIGUOUS]" marker

After the data, provide:
- Total records extracted
- Fields with the highest "NOT_FOUND" rate
- Any patterns or anomalies you noticed in the data

Variables to customize

{{document_type}}{{fields_list}}{{output_format}}{{entity_unit}}

Why this prompt works

Gemini's multimodal document understanding handles PDFs, images of forms, and scanned documents. The explicit normalization rules and NOT_FOUND convention make the output immediately usable in spreadsheets or databases.

Save this prompt to your library

Organize, version, and access your best prompts across ChatGPT, Claude, and Cursor.

Related prompts

Task Decomposition Agent

Forcing the agent to plan before acting prevents premature execution and wasted steps. Explicit dependency mapping enables parallel execution and catches logical gaps early.

Tool Selection Agent

The ReAct pattern (Reason + Act) creates an explicit reasoning trace that improves tool selection accuracy. The error-handling rule prevents infinite retry loops.

Prompt Compressor

Explicitly requiring all functional requirements to be preserved prevents the model from over-compressing and losing critical instructions.

Memory Management Agent

Explicit memory read/write instructions create agents that improve over time. Categorization keeps memories organized, and the deduplication rule prevents context bloat.