How to Use the Data Summarization & Extraction Bot

Our Data Summarization & Extraction Bot is an enterprise-grade extraction engine designed to turn chaotic folders of PDFs and images into structured, actionable data.

Whether you are extracting data from a single receipt to test the system, summarizing 10,000 enterprise invoices, or integrating into your own software, this guide will show you how to format your data, choose your submission method, and retrieve your results.

Step 1: Prepare Your Documents

Your input files should be standard document or image formats (e.g., PDF, JPG, PNG).

Note: The system currently defaults to extracting financial data (invoices/receipts). If you need to summarize other document types (contracts, HR forms, etc.), contact us for a quick custom configuration!

Step 2: Choose Your Submission Method

We support multiple ways to submit your documents, depending on your workflow size and automation needs:

  • Method 1: File Attachment: Best for quick tests and one-off extractions of single documents (PDFs, images, receipts) directly in your browser.
  • Method 2: Cloud Sync (BYOS): Best for automated, bulk processing via Google Drive or Dropbox (delivers full structured data files back to your cloud folder).
  • Method 3: Developer API: Best for large-scale enterprise automation, background processing, and real-time data pipelines.
  • Method 4: Custom Managed Services (Done-For-You): Best for teams with limited technical resources, custom CRM exports, or no time to clean and format raw data.

Review the comparison table below for our self-service browser methods:

Feature & Requirements Method 1: File Attachment Method 2: Cloud Sync (BYOS - Google Drive / Dropbox)
Prompt Input Type any standard instruction OR paste a JSON string for specific custom column extraction. Type any standard instruction OR paste a JSON string to extract specific data across the whole folder.
File Attachment / Storage Setup Upload a single file directly from your computer in the browser. Bring Your Own Storage (BYOS): No manual file uploading. Just input your Google Drive Folder ID or Dropbox Folder Path.
Processing Style Instant, on-screen processing. Automated asynchronous background processing.
Output Delivery Results returned instantly on-screen. Results automatically synced back to your designated Cloud Drive output folder.
Email Notifications N/A Yes. Receive an automated email alert with job details when your cloud folder is populated.

Step 3: Define Your Extraction (Custom Columns)

If you have specific data requirements, you don't have to settle for generic summaries. You can utilize our Custom Columns Configuration in the "Prompt Input" field to tell the AI exactly what data to pull.

Simply provide a JSON string (a simple, standard text format used to store data) of the exact data points you want the AI to find.

Example for Non-Technical Users: If you want the bot to read an invoice and only give you the Vendor Name, Total Amount, and Due Date, you would skip writing a normal sentence and instead paste your text formatted exactly like this into the Prompt Input:

{ "Vendor_Name": "The name of the company issuing the invoice", "Total_Amount": "The final total amount due", "Due_Date": "The date the payment is due" }

The AI will read this list, ignore all the irrelevant text on the document, and output only these three specific columns into your final spreadsheet or JSON response.

Step 4: Run the Engine & Retrieve Your Data

Once you initiate the run, the engine will validate and process your documents based on your chosen method.

If using Method 1 (File Attachment):

The system will process your single file immediately. Your extracted data will be displayed directly on your screen for quick review.

If using Method 2 (Cloud Sync / BYOS) or Method 3 (Developer API):

Because document folders often contain hundreds of heavy files, this system automatically runs in the background so it doesn't freeze your computer or timeout your server.

  • Start the Job: Upon submission, the system generates a unique Job ID (e.g., job_12345). If using the browser UI, you can now close your window and continue working.
  • Monitor Progress (For API/Technical Users): If you are integrating via our API, your developers can ping our /job-status/{job_id} endpoint at any time to see live, real-time metrics (including exact files processed, pages read, and processing speed).
  • Job Completion (For Standard Users): You don't need to monitor the system manually. Once the batch is finished, you will receive an automated Job Completion Notification Email detailing the processing time and status.
  • View Results: Simply open your designated Google Drive or Dropbox output folder to view your newly generated, perfectly structured data files.