Making Haystac(s) Out of Needles

Just Scan the Box of Documents

Excerpts from Document Imaging Report, April 26 2021

We caught up with Haystac CEO Barak Tsivkin to discuss the company’s technology as well as a recent announcement of Haystac’s selection by one of the large information management companies to classify both scanned images as well as electronically stored information (ESI). Haystac is best-known for its content analytics platform used by large enterprises to manage unstructured data at scale, and now their AI-driven technology is transforming the capture market landscape.

Zero companies enjoy paying for the necessary prep work to scan documents. Haystac enables companies to just scan the documents into files – no separator sheets required. “Our Deep Learning algorithms without any training can separate documents that come off the scanner. We can separate documents, tag them, store them, etc.” 

Haystac’s goal is to strip out the complexity of managing unstructured data and place the hard work into the software. That allows non-data scientists to manipulate and extract value from unstructured content.

“We combine unsupervised and supervised machine learning. We use unsupervised machine learning to quickly scan the content. We don’t require any taxonomies or samples prior to this stage. So we can quickly scan the data and put it into large buckets of similar content. Then we use that to feed the machine learning, and we discover the samples and types of categories ourselves.”

“As a result, the data shepherd doesn’t need tell the software what to look for.” This eliminates the need to collect samples to train the software in a classification project, saving months of time. 

The entire article is available to subscribers of Document Imaging Report.  Go here for more information.