Get our latest book on "Top 10 artificial intelligence myths."
Download

Last updated

March 3, 2021

Table of contents

Supported documents

Documents that work well with Impira AutoML

Impira Automated Machine Learning (AutoML) system is built to automate data entry from your documents. Unlike OCR templates, Impira uses machine learning that utilizes geometric information in order to identify the values you want. For example, the location on the page and the proximity to certain nearby words are clues that our system uses to identify the right values. This geometric information allows the AutoML to excel for structured and semi-structured documents such as standard and custom forms, invoices, purchase orders, paystubs, bills, tax documents, and more.

It only takes minutes to extract the information out of your documents and from the first example, and each one after that, Impira AutoML models will learn the specifics of your documents. You can easily extract information and immediately download it to a CSV or export it via our API as part of your workflow.

Forms

Below is a great example of a form for which Impira will excel. Impira AutoML models can extract multiple field types, such as text, dates, numbers, and checkboxes. Further, the AutoML is built to handle the handwriting, rotations, and zooms that you are likely to encounter with scans and images.

Invoices

Impira AutoML natively handles the complexity and diversity inherent to processing invoices for accounts payable or spend analytics. Impira supports field types such as numbers, dates, and text.

Tips for success

For tips on how to get the most of Impira’s AutoML, check out this quick start guide.

Documents that Impira AutoML doesn’t (yet) support

Because Impira AutoML is optimized to use geometric information, there are several types of use cases that our AutoML doesn’t currently support:

  • Extracting specific entities or terms from paragraphs of text that require interpretation.
  • Extracting tabular data.
  • Extracting specific slides from presentations.

We are working hard to add support for more kinds of documents every day. We value any and all feedback about use cases and would love to learn if you are trying to extract something we don’t support today. You can reach us at support@impira.com.

Other ways Impira can work for you

Even for cases where Impira AutoML system can’t currently help with your extraction needs, Impira’s rich functionality can still help automate your workflows. Every file that you upload is available for storage, search, and retrieval. Read more about some tips and tricks for the searching, Impira Query Language (IQL), and integrations here.

All images and documents are run through state-of-the-art OCR models. Even in cases where the AutoML can’t reliably extract your data, users still see dramatic improvements in their efficiency by manually selecting the OCRed text rather than typing it in by hand.