[DigitalToday reporter Seul-gi Son] "A collaboration tool business that only piles up data is no longer viable."
Madras Check CEO Hak-joon Lee (이학준) on June 25 declared at 'AX Festa 2026' held at the Fairmont Ambassador Seoul hotel that the era of collaboration platforms used only by people is over. He said the company will shift its collaboration tool Flow into an artificial intelligence work agent platform. About 1,000 employees, client companies and media representatives attended the event.
Lee cited three reasons SaaS has failed to translate into workplace productivity: a disconnect between data and value, a disconnect between goals and execution, and the limits of people-centric work methods. He presented AI agents as a way to resolve the three disconnects.
Flow’s in-house AI engine, Repattern AI, was also unveiled. It links AI with collaboration and communication, project and schedule, and knowledge and document data accumulated in Flow by about 5,000 client companies over 10 years. Four agents based on Repattern are also provided: Mate, Smart, Consulting and Automation.
Flow also includes a function that assigns an agent as the person in charge of a project in place of a human. Under the system, an agent receives work instructions, asks follow-up questions, updates progress on its own and reports deliverables in comments. Lee said, "When I come to work in the morning and open Flow, the agent has processed 20 to 30 bugs overnight."
It also supports integration with the Model Context Protocol, or MCP. External AI tools such as ChatGPT and Claude can directly call Flow data. The developer center was also revamped so individual employees can receive API keys through an open API and an MCP server and build their own work tools. An internal case was also introduced at the event in which a CRM dashboard was built in three days using Flow MCP and Claude code.
An OKR function that links goals (Plan), execution (Do) and review (See) was also added to Flow. It can automatically calculate goal achievement rates by linking with external performance data from Salesforce, HubSpot and Google Spreadsheets.
Lee said, "The core of Flow’s AI work agent is to amplify value by connecting collaboration and communication data with AI and to connect goals to fast execution."