> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.meetgail.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.meetgail.com/_mcp/server.

# Extract Fields with AI

> Pull structured fields out of an uploaded document, such as a policy PDF.

When a run receives a document like a policy PDF, an ACORD form, or a scanned
application, you often need specific fields out of it rather than the raw text.
The Document Analysis step reads the document and returns the fields you ask for
as clean, structured values.

## Nodes involved

* [Document Analysis](/platform/workflows/node-reference/ai/document-analysis) reads a
  document and extracts the fields you describe.

## How to build it

### Get the document into the run

A document usually arrives as a file input when the workflow starts, or from an
earlier step that produced or downloaded a file. See
[Starting a Workflow](/platform/workflows/core-concepts/starting-a-workflow) for how
file inputs work.

### Describe the fields you want

Add a Document Analysis step, point it at the document, and describe the fields
to pull out, such as the policyholder name, policy number, and effective date.
The step returns those fields as structured values.

### Use the extracted fields

Later steps can read each field by name and act on it, for example saving it to
a contact or routing the run based on its value.

Ask for exactly the fields you need, named clearly. A focused list gives more
reliable results than asking for everything at once.

## Related

#### [Document Analysis](/platform/workflows/node-reference/ai/document-analysis)

Every option for extracting fields from a document.

#### [Starting a Workflow](/platform/workflows/core-concepts/starting-a-workflow)

How a document gets into a run as a file input.