Connexioh

ConnexioH (코넥시오에이치, also referenced as CAH) is a South Korean data intelligence company that collects and refines e-commerce data and supplies it to financial institutions, card companies and brands. Its stack combines an ontology layer that structures fragmented e-commerce business data into a knowledge graph, a Graph RAG-based commerce-specific AI model intended to suppress LLM hallucination, and an MCP/API streaming layer for real-time internal and external data delivery.

The product line spans fintech and commerce. 'CH DataLake' is a data specification platform that processes e-commerce data into a form usable for alternative credit scoring and delivers it periodically; the company supplies more than 110 data items monthly to NICE Information Service. 'CAH Cre-pan (On-sell)' is a visualization tool for bank branches that maps an online seller's sales flow, location, business status and growth potential to support credit assessment, and Cre-pan scores sellers using non-financial data such as sales history, inventory turnover and review reliability, with direct API connections to financial institutions. 'Memopan' is an AI analytics dashboard for Naver Smart Store sellers, focused on search engine optimization support by analyzing product exposure, search keywords, and review and 'like' trends. A Commerce AI Agent runs on the ontology engine and automates pricing strategy, inventory operations and competitor monitoring with a natural-language query interface, demand forecasting and multi-channel dashboards.

Founding story

CEO Kyung-ho Lee worked as an IT security expert before becoming a professor at Korea University, where he ran a risk management lab studying corporate credit risk and found that data gaps and contextual understanding of unstructured data were central to identifying risk. Observing that small merchants who moved to e-commerce could not obtain financing because credit assessment relied on periodic financial statements or physical collateral, he concluded that real-time online platform activity data could support fairer evaluation. He established the company through a faculty-led startup program, positioning it as a link between risk-averse lenders and online sellers seeking to evidence their performance.

Business model

B2B data supply and software. ConnexioH sells processed e-commerce data and credit-scoring inputs to banks, card companies and credit bureaus (including periodic delivery of over 110 data items to NICE Information Service and supply contracts with KB Kookmin Card and Shinhan Card), provides visualization tooling to bank branches, and offers seller-facing analytics through Memopan.

Traction

The website cites over 100.7 billion accumulated product data records, 2.11 million businesses analyzed and a data valuation figure of 609 million. Press reporting cites analysis of data from over 1.18 million businesses, more than 300 million data-driven products and over 3 million collected keywords. Memopan signed up 10,000 businesses in its first week and has surpassed 50,000 cumulative users, addressing roughly 570,000 Naver Smart Store sellers. The company reports completed financial-sector data supply contracts with Kakao Bank, KB Kookmin Bank and Shinhan Card, supply of over 110 monthly data items to NICE Information Service, and FSC-selected pilot testing with BNK Busan Bank.

Latest developments

In March 2025 the company disclosed investment via the CNTTech DB Dream Big Fund, jointly managed by CNTTech and DB Capital, following legal and technical due diligence. In April 2025 it was selected as an official provider for a Financial Services Commission commissioned test, and by October 2025 reported completed financial-sector data supply contracts with institutions including Kakao Bank, KB Kookmin Bank and Shinhan Card. An October 2025 interview described On-sell testing with BNK Busan Bank and a planned full release of Memopan to all users in the following year.