Korea Deep Learning (KDL), an enterprise artificial intelligence agent firm, is scaling its footprint across the Asia-Pacific financial sector following a successful proof of concept (PoC) alongside South Korea's Industrial Bank of Korea (IBK). According to the announcement, the collaboration yielded an 88% decrease in financial document processing time.
The deployment utilized KDL's vision-language model (VLM)-based artificial intelligence agent known as DEEP Agent for Finance. The system was designed to automate document-heavy workflows associated with processing voice phishing relief cases. By integrating document recognition, information extraction, and crime-type inference, the technology tackles complex assignments that traditionally demand extensive manual evaluation by bank personnel.
Evaluation metrics from the project indicated that the platform attained 100% accuracy in printed-text recognition and document categorization. Furthermore, handwriting recognition accuracy surpassed 97% under the evaluation criteria, allowing all key performance indicators to surpass initial objectives using genuine business data. The deployment occurred securely inside IBK's closed network environment, maintaining human employee review safeguards for scenarios requiring supplementary verification.
This implementation serves as a crucial reference point as KDL targets broader international expansion. The enterprise previously finalized a separate PoC with Shinhan Vietnam Finance Company (SVFC), an operating subsidiary of Shinhan Card, via the Shinhan Financial Group open innovation initiative. Additionally, KDL has deployed artificial intelligence solutions in foreign markets such as Vietnam and the Taiwan region.
Looking toward Japan, KDL is currently progressing discussions regarding prospective PoCs with nine separate financial institutions. The firm is additionally maintaining engagements with banking, insurance, and consumer finance entities throughout the region. Certain prospective partners have already executed non-disclosure agreements and are actively assessing the technology using real operational data.
Moving forward into 2027, KDL intends to transition current technical assessments into paid proofs of concept and formal supply contracts, concurrently building local alliances and distribution channels. The organization also plans to broaden its offerings outside traditional banking into the insurance and securities sectors, where information extraction and document handling remain vital daily functions.
Established in 2019, Korea Deep Learning focuses on document intelligence and automated workflows through its proprietary DEEP Agent platform, which merges DEEP OCR and DEEP Parser solutions to retrieve and interpret data from intricate documentation. The business currently provides services to over 80 corporate clients and public agencies within South Korea.
