Cogent Labs

Cogent Labs's valuation is not publicly disclosed.

Investor information for Cogent Labs is not publicly available at this time.

No public funding round data is currently available for Cogent Labs.

Cogent Labs has approximately 100-500 employees.

Cogent Labs develops AI document processing and text analysis software with specialised capabilities in Japanese character recognition, handwritten text, and structured document data extraction, targeting the Japanese financial services, insurance, and government sectors where large volumes of handwritten and mixed-format documents require processing with high accuracy in Japanese scripts including kanji, hiragana, and katakana. The Tokyo company applies deep learning models trained on large Japanese document datasets to achieve recognition accuracy on complex Japanese text that general OCR systems cannot match.\n\nThe company raised approximately $50 million in venture funding from investors including WiL and Goldman Sachs. Cogent Labs Tegaki handwriting recognition product has been deployed by major Japanese insurance companies, financial institutions, and public sector organisations to automate document digitisation workflows that previously required large manual data entry teams. The Japanese market for document AI is substantial given the volume of paper-based documents in government and financial services operations and the complexity of Japanese script that makes off-the-shelf OCR insufficient.\n\nCogent Labs competes in the Japanese document AI market against OBIC, NTT Data, and international intelligent document processing vendors including ABBYY and Kofax. Its Japanese-specific technical capabilities create a natural market advantage that English-first vendors struggle to match through localisation alone, as accurate Japanese handwriting recognition requires specialised model training that cannot be derived from models built primarily on Latin character datasets. The company has expanded internationally to address Korean and other East Asian scripts with similar character recognition complexity.