5.0

Release Date: December 17, 2025

Canopy is excited to introduce Auto Review, the first leading Agentic AI review software designed to revolutionize your document review workflow.

What it is: Auto Review is an Agentic AI review engine that acts as the initial, first-level document reviewer. Auto Review uses AI and user-input Entity Qualifying Rules to automatically identify and propagate entities directly to the documents.

How it helps: Auto Review streamlines your document review process by eliminating the need for manual human review of large, structured document sets. This automation dramatically increases efficiency and reduces costs.

Auto Review streamlines your review process, saving you time and delivering unprecedented efficiency for your document review projects.

For more details on this feature, refer to the Auto Review documentation.

We have removed the previous batching restrictions to provide you with unparalleled flexibility in your review process.

Previous Batching

In the past, users were required to follow a specific review workflow, such as Review -> QA. Additionally, a document could only belong to one batch at a time.

New Batching

Now, there are no workflow restrictions, and you can batch documents more freely. Furthermore, a single document can now belong to multiple Review Phases (Batch Sets) simultaneously.

This enhancement lets you create a flexible review workflow tailored to your team’s and project’s specific needs.

For more details on this feature, refer to the Review Phase documentation.

We have deployed a dedicated OCR service that standardizes document processing on a unified technology stack, delivering greater consistency and accuracy across the board. Key enhancements include:

• Expanded Language Support: Robust text extraction for Chinese, Japanese, and Latin scripts, including improved handling of standard Latin characters (not including diacritics).
• Advanced Handwriting Recognition: Significant improvements in identifying and extracting data from handwritten entries.
• Intelligent Image Handling: Improved accuracy for images requiring rotation or orientation adjustment.
• Increased Reliability: Enhanced process resiliency and higher extraction confidence scores across all document types.

Enhanced the system’s capability to process a diverse range of encoding formats, ensuring fewer document errors during upload and processing.

Resolved an issue in the ZIP processor where files incorrectly inherited hashes from their parent archive, leading to duplicate document IDs for files within the same package. All affected projects have been identified and programmatically repaired to ensure unique document identification.