Sansan technology powers
AX (AI transformation)
Generative AI only delivers real value in business when accurate business data is in place. Sansan has the proprietary technology to accurately digitize analog information such as business cards and contacts, invoices, and contracts.
Proprietary digitization
We’ve built a proprietary digitization system that combines machine learning, generative AI, and human input to deliver accuracy, speed, and security. We’ve developed our own generative AI on a foundation of technology and expertise refined over many years, further advancing process efficiency.
Proprietary image recognition technology
We digitize with high speed and accuracy through a range of internally developed image recognition technologies.
NineOCR
NineOCR is our OCR engine purpose-built for business cards, pairing a detector that identifies field attributes on each card with a recognizer that accurately reads names and other expressions specific to business cards. With NineOCR, we’ve resolved legacy OCR’s accuracy challenges, reaching a high degree of accuracy while cutting digitization costs.
Viola
Viola is our generative AI built on a Vision Language Model trained on a wide range of document data. We turn domain-specific models tuned to digitization rules for each business area into APIs, embed them into our systems, and run them as digitization engines.
Cello
The Cello AI model adds a position-output mechanism to Viola. With position output, we can extract information from fine text through integration with NineOCR and other tools, heightening digitization accuracy. We develop this model with support from GENIAC (Generative AI Accelerator Challenge), a project run by Ministry of Economy, Trade and Industry of Japan (METI) and the New Energy and Industrial Technology Development Organization (NEDO), to strengthen domestic generative AI development capabilities.
Watch this video for a clear explanation
CTO’s message
Creating reliable data that
drives business
Sansan’s mission is “Turning encounters into innovation,” as we reshape the commonly accepted ways of doing things in business. We turn the primary information born from those encounters – including business cards and contacts, invoices, and contracts – into data that businesses can use. That is how we drive innovation. We’ve spent years refining our system that combines machine learning, generative AI, and human-driven quality control, all built on a foundation of digitization rules. We’ve developed our own image processing and natural language processing technologies, and we remain strongly committed to producing precisely structured data. Business AI’s inference accuracy ties directly to the underlying data’s quality. Unstructured data mixed with noise is what causes AI to make critical misreadings. The technical prerequisite for making AI work in real operations is preparing structured, factual data that strips out guesswork and can be read accurately. This operational discipline of carrying accurate digitization and quality control through to the end is precisely what makes our value distinct. The technologies we’ve cultivated are what enable our AX services that reshape how we work.
Executive Officer, CTO
Hirohito Sasakawa
