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.

Thank you for read­ing this post, don’t for­get to sub­scribe!

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.