Content
What’s new? Improve digitization accuracy with OCR confidence  Add third-party data, calculations, and more with custom function fields Automatically process email attachments Enhanced barcode and QR code support Support large-scale use cases with larger allowances for fields, classes, and input files Additional improvements Try out AI Hub today

Over this past month, AI Hub has continued to evolve and enhance its capabilities. This month’s update brings exciting new features designed to:

  • Streamline your workflow
  • Improve accuracy
  • Support more complex use cases. 

From automatic processing of email attachments to powerful custom function fields and expanded support for large-scale projects, we’re committed to making your experience with AI Hub more efficient and robust. Watch the following video to learn more.

Let’s take a deeper dive into the latest additions and improvements.

Converting images into machine-readable text is the first step to extracting information from documents. However, handwritten or blurry documents often pose accuracy challenges. To address this, AI Hub now includes OCR confidence scores to validate digitized text. OCR confidence measures how reliable the model is about the accuracy of the transcribed information. A high confidence score indicates high accuracy, while a lower score suggests that the data might not be entirely precise. This feature allows users to gauge the reliability of digitized text against validation thresholds, flagging low confidence outputs for human review and improving the overall accuracy of extracted information

AI Hub enhances its functionality with the introduction of custom function fields. These fields allow you to enrich document results with third-party data, perform calculations, and incorporate custom business logic. You can now use Python logic to create custom fields that compute values or import data into your projects. For example, you can retrieve additional information from third-party systems, append customer data from your CRM, or calculate financial metrics like debt-to-equity ratios based on extracted data. This capability provides greater flexibility and ensures that the data you work with is both comprehensive and tailored to your specific needs.

Gone are the days of manually sifting through emails and their attachments to find specific documents. While AI Hub previously could process email files, handling attachments required additional configuration. With the latest update, AI Hub now automates this process. When you upload an .eml email file, AI Hub seamlessly processes the email and all its attached documents. Each attachment is individually split and added to your company’s knowledge base, simplifying document management and saving valuable time.

Barcodes streamline data entry and are often used to classify documents, provide unique identifiers, and encode additional information. By default, AI Hub automatically detects and decodes barcodes and QR codes. This release improves the reliability of extracting and handling barcode information when using prompts. 

AI Hub now supports more extensive and complex use cases with increased allowances for fields, classes, and input files. You can now utilize up to 100 fields per class and manage 250 classes per project. This enhancement is particularly beneficial for projects involving diverse document types or those requiring extensive field extractions. 

Additionally, AI Hub supports up to 500 input files in the Build phase, which enables better configuration for documents with high variability. This feature allows for more rigorous testing and refinement of extraction prompts, ensuring robust and accurate solutions for large-scale projects.

  • Autosave: Classes and fields are now autosaved as you work, simplifying the configuration process and ensuring changes are saved should anything unexpected occur.
  • Faster Extractions: AI Hub’s extraction rate speeds are now roughly 50% faster than previous versions. This upgrade translates to quicker processing times and enhanced user productivity, allowing you to accomplish more in less time.
  • Automatically Switch Between Models: AI Hub now offers the ability to automatically switch between models based on the specific needs of the task, such as document length and quantity. This feature eliminates the need for manual model selection, as AI Hub will assign the most suitable model for the job, optimizing performance and efficiency.
  • Date/Time Awareness: AI Hub now includes date and time awareness. Date and time fields will automatically assume the current date and time unless specified otherwise in your prompt. This enhancement simplifies tracking and managing document extractions, making your workflow more intuitive.

With these powerful updates, AI Hub is poised to enhance your team’s productivity, accuracy, and scalability like never before. Whether you’re looking to streamline document processing, enrich your data, or tackle large-scale projects, these new features are designed to support your goals effectively. We encourage you to explore these enhancements and see how they can transform your workflow. Try AI Hub for free today.

Explore the latest AI Hub updates

See how Instabase makes it easy to accurately extract data from unstructured documents.