HSAD took part in "Braincrew Connect 2026" to share AI transformation (AX) cases in the advertising and marketing industry and unveiled the AI solution "Deep Agent Builder" for the first time.
On the 14th at Maru180 in Seoul, Braincrew Connect 2026 was a conference hosted by Braincrew, a partner of the global AI orchestration platform LangChain. Domestic AI experts, including Choi Hoon, chief development officer at Upstage, which is participating in a national AI project, and Jang Dongin, professor at KAIST Graduate School of AI, joined as speakers. HSAD was the only participant from the advertising and marketing industry.
Last year, HSAD became the first in the advertising industry to build "Dashflow," an open agent platform that lets working-level staff design and use AI agents directly. Without separate coding, HSAD members have directly created AI agents needed for their work by consolidating agent builders such as Dify and n8n and designing the execution order, and built an agent ecosystem shared by teams, business divisions, and the entire company by consolidating them into Dashflow.
That day, HSAD gave an early reveal of "Deep Agent Builder," Korea's first natural language–based agent builder co-developed with Braincrew. Braincrew CEO Lee Kyung-rok said, "Deep Agent Builder makes it easy to build AI agents tailored to each person's job objective with natural language conversations alone."
For example, when a user states requirements in everyday language such as "I want to create a web search agent" or "Create an agent that recommends optimized copy by media," the solution can grasp the user's intent, recommend the optimal system prompts and tools, and generate the agent. The agent builder is scheduled to be loaded into Dashflow at the end of February.
Lee Sung-jae, HSAD chief strategy officer (CSO), said, "With Deep Agent Builder, even marketers without development knowledge can implement AI agents if they have creative planning skills," adding, "We expect members will be able to freely create and use agents that are useful for project-specific tasks."