Course Content
Flows – Master
0/7
Working with Files – AI Agent
00:00
DocRag AI – AI Agent
00:00
Formatting – AI Agent
00:00
The Schedule Trigger – AI Agent
00:00
Extract Document Table – AI Agent
00:00
Webhooks – AI Agent
00:00
Course Exercise: Create a new exercise with Doc rag – AI Agent
00:00
Flows – Master

Lesson Guide: Intelligent Document Processing with DocRag AI

In this lesson, you will learn how to read, interpret, and extract critical information from unstructured files using Noca’s native DocRag AI node.

Manually downloading PDFs or images (like W-9 forms, invoices, or contracts) to copy-paste names, addresses, or tax IDs into your database is slow and prone to errors. DocRag AI eliminates this manual data entry. By placing it directly on your canvas path, Noca automatically parses complex document layouts into readable data layers, allowing you to extract structural text arrays using simple, plain-English prompts.

Step 1: Establish Your File Source Path

Before the AI engine can analyze a document, you must first direct it to the physical file payload.

  1. Open your Noca Flow workspace and navigate to the canvas path configured in the previous lesson.

  2. Confirm your workflow sequence flows correctly from Salesforce Source (Related Files).

  3. Click the + icon located immediately downstream from your Related Files source block.

Step 2: Insert the DocRag AI Node

With your unstructured attachment stream established, call upon Noca’s intelligent processing layer.

  1. In the node action selection panel that slides out, type DocRag AI into the search box.

  2. Select the DocRag AI node and click Insert Node to drop it directly onto your execution timeline.

Step 3: Map Your File Asset and Design Prompt Queries

Now, feed your physical document variable into the AI node and define exactly what data points you want to extract.

  1. Click on the newly placed DocRag AI Node to reveal its configuration properties on the right.

  2. In the file or object input slot, open your left-hand Data Kit explorer and drag the file payload reference variable from your Related Files node straight into the target input box.

  3. Locate the Prompt / Query configuration area.

  4. Instead of writing code, simply type out your data extraction requirements in plain text. For example:

    “Extract the Taxpayer Identification Number (TIN), business type, and the full address listed on this W-9 document. Return the response in a clean structured format.”

Step 4: Map Your Extracted Output Fields

Once DocRag AI finishes analyzing the layout, it isolates your prompted targets into clean variables.

  1. Add your next downstream action step (e.g., a Salesforce Update Record node or an internal spreadsheet row creator).

  2. Open your left-hand Data Kit and open the dropdown menu next to your DocRag AI node.

  3. Result: You will find perfectly extracted text fields ready to be dragged and dropped into your final data fields—allowing the system to automatically populate structural CRM variables straight from an unreadable PDF or image!

Key Takeaways

  • Automated Layout Parsing: DocRag AI takes unstructured images or document attachments and autonomously splits them into machine-readable text arrays.

  • Prompt-Driven Extraction: You do not need to configure rigid coordinate bounding boxes or regex lines. Simply describe what information you want in plain natural language, and the AI will pull it out.

Fantastic work! You have successfully mastered unstructured intelligent document processing. You are now fully equipped to build highly adaptive, end-to-end automated pipelines across the Noca Flow ecosystem!

Previous
Next