Ask AI
How can we help? 👋

Sonar Beta Program: AI Data Extraction from Files

What is Sonar?

Sonar Beta is our first AI-powered feature for document management. We built it to help you capture more data in ORCA, faster and with less risk of human error. The feature automatically extracts key information from your files, builds your ORCA structure, and streamlines the way you process and file documents.

Notion image

How it works

With Sonar, information from your uploaded documents is automatically analyzed by AI and transformed into suggested events. You can review and add these with a click, making it simpler and faster to log details and keep your documents organized.

A simple 3-step process:

  1. Upload: Add documents to ORCA; Sonar AI analyzes them.
  1. Review: Verify the AI-suggested events and entities; edit or add details if needed.
  1. Save: Approve to update your ORCA structure and file the document.

Example: Upload a purchase agreement → Sonar proposes: “Horizon Capital AG bought 1,000 shares of Riverton Labs Inc. for USD 1,200,000 on 15.03.2024.” → You confirm and save. The event is added to existing entity Horizon Capital AG and a new entity is created for Riverton Labs Inc.

Extracted information

Sonar analyses the content of documents to extract key legal and corporate information. This information currently includes:

  • Ownership/Corporate structures and Transactions: purchases, ownership stakes, transfers, and share classes.
  • Identities: names and addresses of legal entities, assets and persons.
  • Roles and Mandates: powers of attorney, general partnerships, limited partnership.
  • Loans.
  • Document records: passports held, articles of association, certificates of incorporation.
 

Sonar isn't limited to specific document types. If your document contains relevant information, Sonar will attempt to extract it.

The following document types typically contain the most relevant information:

  • Ownership Charts (from PowerPoint, Visio, or napkin sketches; just export or capture it as a PDF or image before uploading)
  • Cap Tables
  • Purchase Agreements
  • Articles of Association
  • Certificate of Incorporation
  • Limited Partnership Agreements
  • Powers of Attorney
  • Loan Agreements
  • Passports

Running Sonar on Existing Vault Documents

To help you identify unprocessed files, the system automatically populates a Suggested for Sonar folder.

  • Criteria: This folder captures any document uploaded within the last 90 days that is not yet linked to an Event.
  • Automatic Maintenance: Once a document is reviewed or successfully linked to an Event, it is automatically removed from this folder to keep your view organized.

This works especially well for files uploaded via e-mail to your ORCA vault.

You can process multiple files simultaneously from either the Files main section or the Entity Details View.

  1. Navigate to the Files tab or a specific Entity.
  1. Open the Suggested for Sonar folder.
  1. Select the checkboxes next to the documents you wish to scan (you can select up to 10 documents at a time).
  1. Click the Run Sonar button in the action toolbar.
Notion image

Rerun Sonar Analysis

Rerun Sonar analysis on a file directly from the review screen. The rerun triggers a fresh analysis using the existing file and replaces the current extraction results, making it easier to quickly retry and improve AI analysis outcomes.

Notion image

Bulk Delete Events

Select and delete multiple events simultaneously directly from the File Review window, eliminating the need to remove them one by one.

Notion image

Intelligent Event Matching

To ensure your data remains clean and organized, Sonar now automatically checks your existing Vault data before suggesting new entries. This reconciliation process prevents data fragmentation and saves you from manual cleanup.

  • Smart Reconciliation: If Sonar extracts an event (such as a share purchase or a change of address) that already exists in your Vault, it will intelligently link the information to the existing record rather than creating a duplicate entry.
  • Visual Indicators: The review interface now features distinct UI indicators to help you distinguish between entirely new data points and those that have been matched with existing events.
    • Notion image
  • Purple Highlights: When Sonar identifies new details for an existing event, these specific fields are highlighted in purple. This allows you to review updates or changes at a glance without having to re-read the entire entry.
    • Notion image

Key details

  • Formats: PDF and images.
  • Browser: Chrome or Edge required.
  • Processing and reviewing: In-session only. AI results are visible only to the uploader until reviewed and saved.

Beta program

  • Pricing: Free during beta.

Privacy & data security

  • Zero data retention.
  • Files never used for AI training.
  • Secure OpenAI integration.
  • EU-based processing.
  • Files are briefly exposed to ORCA servers during processing only.

Interested?

Enable Sonar in Account Settings → AI features. See more information here.

 
Did this answer your question?
😞
😐
🤩