Cognitive Clas­si­fi­ca­tion for Input Manage­ment

Today, companies receive infor­ma­tion in a wide variety of formats and forms, including text, images, tables and forms. Often, these formats are mixed and this is also pur­po­seful, because a piece of infor­ma­tion can be expressed dif­fer­ently and better in a certain format than in others.

For auto­ma­tion tools, on the other hand, this hete­ro­ge­neous flow of infor­ma­tion often poses a major challenge. Classic clas­si­fiers in input manage­ment are often optimized for specific formats and therefore quickly reach their limits. Not so the new cognitive clas­si­fier from Insiders, which relies on Deep Learning – and thus takes the auto­ma­tion of business processes to a new level.

„The more precise and com­pre­hen­sive the input manage­ment, the more effi­ci­ently the data can be used in the sub­se­quent process flow.“

The hybrid cognitive clas­si­fier developed by Insiders Tech­no­lo­gies analyzes a document as a whole in one step, combining cognitive capa­bi­li­ties such as reco­gni­zing images, reading, seeing struc­tures and layouts. The hybrid cognitive clas­si­fier inte­grates different AI methods and merges their respec­tive strengths into a common deep learning algorithm. This com­bi­na­tion in par­ti­cular results in signi­fi­cant process acce­le­ra­tions and increased auto­ma­tion rates.

In its current issue (76th volume | November 2021), Ver­si­che­rungs­wirt­schaft also reports on the cognitive clas­si­fier from Insiders Tech­no­lo­gies. Using practical examples from the insurance industry, authors Karl-Josef Krechel-Mohr and Dr. Darko Obradovic demons­trate the enormous potential savings that these new clas­si­fi­ca­tion options unleash in companies‘ input manage­ment.