Using Vision AI to extract data from complex documents

Ruth Yakubu

Challenges of Processing unstructured files

We’ve all have experienced the frustration of an application or payment being delayed due to paperwork. This delay is usually due to several factors involved processing of documents. Businesses have to process massive amounts of data that are structured or unstructured. Most of this data are files. These files can be passports, invoices, bank statements, receipts, license plates, forms and more. These contents must be digitally stored, processed and be searchable. However, documents have its own complexities, they have both printed and handwritten text, images, language, barcodes, and other characters. In addition, the documents come in different file formats like TXT, DOC, PDF, JPEG, BMP and TIFF. These complexities make businesses operate at a slower pace by manually processing the data or searching for dependent information. These are all factors that have a huge impact on customer satisfaction and challenges companies face in document processing.

Driving Innovation

In this episode of Tech Exceptions, I had an exceptional conversation with Anant Bhardwaj, CEO of Instabase, and Kerry Chang, Tech Lead at Instabase who have raise over $131M in funding for their Al platform designed to make computers work for people, and empowering people to focus on the meaningful and productive parts of their job. They are solving this problem by boosting content discoverability and extracting data trapped inside complex documents with unmatched quality. Instabase takes it a step further by enabling organizations to build workflows to access information in files and automate business processes. Regardless of the input file format or desired end-result, their solution helps businesses handle the most complex file content processing workflows.

Leveraging Azure AI

We discussed how, unicorn startup, Instabase is using Azure Computer Vision which includes Optical Character Recognition (OCR) capabilities to extract data from documents or images. The OCR supports extracting printed and handwritten text from images and documents; mixed languages; digits; currency symbols and writing styles to help businesses to process documents more efficiently and improve customer experience. We also discussed interesting use cases where their use of Azure Computer Vision’s Read Docker container gives them the option to offer on-premise support to enterprises with security and data governance restrictions to keep their data off the cloud.

The Solution

Watch and learn how Instabase is using AI to transform how enterprises work with complex documents and improve productivity with their AI platform using Computer Vision and Natural Language Processing (NLP) to deliver exceptional automated customer experiences.

Coming up Next…

Mark your calendars! I’ll be joined by Nina Alag Suri, CEO of X0PA AI to learn how the company is using Cognitive Services, NLP and Bots in their AI solution to eliminate hiring bias by providing powerful pre-screening and predictive insights to recruiters and hiring managers so they can make more accurate best fit selection decisions to enhance the quality of talents and an inclusive workforce for companies. Watch Live on Learn TV March 10th 2:00-2:30pm PT!

Ruth Yakubu Senior Cloud Advocate, Developer Relations

Follow