AI-powered clas­si­fi­ca­tion boosts auto­ma­tion

Using the Cognitive Clas­si­fier from Insiders Tech­no­lo­gies, BGV has increased full auto­ma­tion in input manage­ment by 10 percent. With more than 650,000 incoming documents annually and an average clas­si­fi­ca­tion accuracy of around 95 percent, BGV has achieved a new level of quality in automated pro­ces­sing, notably reducing employee workload and ensuring stable processes even during peak periods.

The BGV Badische Ver­si­che­rungen Insurance Group offers com­pre­hen­sive insurance coverage for muni­ci­pa­li­ties, busi­nesses, and private customers. Deeply rooted in the Baden region of Germany, BGV stands for regional respon­si­bi­lity, part­ner­ship-based col­la­bo­ra­tion, and sus­tainable solutions. At the same time, demands for effi­ci­ency, trans­pa­rency, and digital processes are growing.

To manage complex business processes cost-effec­tively and reliably meet regu­la­tory requi­re­ments, input manage­ment is incre­asingly taking center stage.

BGV processes more than 650,000 documents each year in its central inbox. On regular business days, an average of around 3,000 items is received, with signi­fi­cant seasonal peaks. The vast majority reach the company digitally via email, with postal mail as a sup­ple­ment. The range of content is enormous: claims reports, contract amend­ments, can­cel­la­tions, inquiries, or indi­vi­dual customer concerns. In total, the insurer distin­gu­ishes between appro­xi­m­ately 110 different business tran­sac­tions and routes them to the respec­tive spe­cia­lized processes.

DIVERSE INBOUND CHANNELS AND HIGH-QUALITY REQUI­RE­MENTS

With incre­asing digi­ta­liza­tion, the structure of incoming documents has changed signi­fi­cantly. Different layouts, varying amounts of text, embedded images, or attach­ments con­sider­ably increase the com­ple­xity of clas­si­fi­ca­tion. At the same time, regu­la­tory requi­re­ments, service level agree­ments, and internal quality standards remain con­sis­t­ently high.

For BGV, input manage­ment is therefore not an isolated pre-sorting step, but a central component of the entire process chain. Clas­si­fi­ca­tion must be robust from the outset to reliably trigger sub­se­quent processes and ensure trans­pa­rency regarding the overall pro­ces­sing status.

Nutzen für den Kunden

  • Stei­ge­rung der Dun­kel­ver­ar­bei­tungs­quote im Post­ein­gang um 10 Pro­zent­punkte auf 95 Prozent
  • Deutlich redu­zierte manuelle Nach­be­ar­bei­tung
  • Höhere Präzision durch kom­bi­nierte Bild- und Tex­terken­nung

„The new Cognitive Clas­si­fier is a signi­fi­cant step in our long-standing col­la­bo­ra­tion with Insiders. It allows us to gradually integrate AI into our existing systems and achieve immediate, reliable results. This signi­fi­cantly reduces the effort required for cla­ri­fi­ca­tion and has nearly doubled our pro­duc­ti­vity in post-pro­ces­sing.“

Sören Hauser

Team­leiter Input Manage­ment, BGV

chal­lenges

  • Zuneh­mende Vielfalt und Kom­ple­xität ein­ge­hender Dokumente
  • Hohe Volumina mit sai­so­nalen Spitzen
  • Wunsch nach höherer Dun­kel­ver­ar­bei­tung bei stabilen Prozessen

„The key challenge is the scalable, stable, and con­troll­able pro­ces­sing of our highly hete­ro­ge­neous incoming customer cor­re­spon­dence – with high volumes and clear quality requi­re­ments.”

Sören Hauser

Team­leiter Input Manage­ment, BGV

CON­TI­NUITY IN COL­LA­BO­RA­TION AND DELI­BE­RATE FURTHER DEVE­LO­P­MENT

For several years now, BGV has relied on the smart FIX solution from Insiders Tech­no­lo­gies for input manage­ment, now stra­te­gi­cally enhanced as part of the OmnIA platform. In addition to its tech­no­lo­gical capa­bi­li­ties, BGV was par­ti­cu­larly impressed by the company’s proven industry expertise in the insurance sector and the ability to develop the platform incre­men­tally.

To further enhance the quality of its existing auto­ma­tion, BGV sup­ple­mented the estab­lished solution with Insiders’ Cognitive Clas­si­fier as the next logical step in its deve­lo­p­ment. This AI component analyzes text and image elements using spe­cia­lized methods for each, thereby signi­fi­cantly improving clas­si­fi­ca­tion within the estab­lished process archi­tec­ture.

The tran­si­tion was deli­bera­tely carried out during a period of increased workload asso­ciated with the annual financial state­ments – intern­ally described as “open-heart surgery” – to quickly benefit from the expected impro­ve­ments. Struc­tured pre­pa­ra­tion, close moni­to­ring, the stability of the existing system landscape, and con­fi­dence in the solution and the project team ensured a suc­cessful tran­si­tion to pro­duc­tion.

Read the full success story now

Name
Anmeldung zum News­letter