Company Overview#
StepHow is a Korean company that builds AI products for enterprise knowledge and data. Its flagship product, Wissly, is an AI system that connects a company’s internal data and answers questions against it. The company also runs a US subsidiary in Delaware for overseas expansion.
StepHow was founded in 2024 and is based in Seoul. The team is small and focused on enterprise AI and large-scale document processing. Alongside Wissly, the company offers StepHow for AI-generated work manuals and knowledge management.
Technology & Product#
Wissly connects documents, shared folders, NAS, ERP, and groupware into one AI workspace. Users ask questions in plain language, find the right information across large document sets, and check the original source passages next to each answer. They can then turn the result into summaries, comparison tables, and report drafts.
The core is a document processing pipeline built for scale. Wissly parses, chunks, embeds, and indexes files, including scanned PDFs and tables through OCR, so it can search by meaning rather than by filename. Search combines vector, keyword, and graph-based methods, and every answer is backed by cited sources. Access follows each organization’s data structure and permissions.
Security is central to the design. Wissly can run fully on-premise through an integrated appliance, Wissly-in-a-Box, for organizations that cannot send internal data to an external cloud. Two related products extend the platform: Ontology defines shared meanings and relationships across company data, and Wissly Code runs AI coding on an internal network using a local LLM.
Market & Use Cases#
StepHow targets enterprises and public institutions that hold large volumes of scattered internal data and face strict security requirements. Typical early adopters are in construction, manufacturing, and the public sector, where contracts, drawings, specifications, schedules, and approval documents pile up across many systems.
Common use cases include cross-checking contracts and technical documents, answering internal process questions such as what forms a travel expense claim needs, analyzing spreadsheets, and drafting reports from verified sources. Because answers link back to the original documents, the tool fits teams that need to trust and audit the result, not just read a quick summary.
Traction & References#
Wissly is in use and in active pilots with enterprise and public-sector customers. The City of Tallinn in Estonia ran a pilot for source-backed search across public administration documents, letting users verify answers against the original policies(Wissly). Hoban Construction is validating document AI across contracts, schedules, and technical reviews(Wissly), and Techcross is another customer(Wissly). Other customers include KAIST, Pulmuone, and Stemco(Wissly).
A product walkthrough covering document search, source checking, spreadsheet analysis, and report drafting is available as a demo video(YouTube).
Collaboration Relevance#
StepHow fits enterprises, financial institutions, and public agencies that want AI over their own data but cannot move that data to a public cloud. A natural PoC connects a document, NAS, or ERP corpus and tests source-backed search and report drafting against real internal content, with deployment scoped to the customer’s security review.
For Korea-Europe collaboration, the City of Tallinn pilot shows early European public-sector traction. Wissly’s on-premise option is relevant for regulated and multilingual European organizations that need reliable process documentation and auditable document search. The Delaware subsidiary supports parallel expansion into the US market.