AI services and solutions company Bespin Global will hold an "AX breakfast seminar for manufacturing company CxOs" on Oct. 28 at Josun Palace Seoul Gangnam under the theme, "Practical AI adoption strategies seen through AI transformation cases of domestic and overseas manufacturers."

At the event, Bespin Global will introduce its Journey to AI diagnostic methodology, which measures each company’s current position and baseline figures over 6 to 8 weeks. It will also present an execution model in which a forward deployed engineer (FDE) takes a single task from diagnosis to financial verification.

Bespin Global will also unveil a manufacturing AX strategy map based on its "7 Layer AI Stack," systematised through experience operating infrastructure for about 5,000 client companies and carrying out more than 200 data and AI projects.

The 7 Layer AI Stack consists of 7 layers: infrastructure, model and LLMOps, DataOps, ontology, agent build and operations, governance and security, and performance management.

For manufacturing applications, it presents global factory data standardisation (L3 DataOps), turning equipment and quality data into a knowledge graph (L4 ontology), building and operating AI agents on production sites (L5 agent build and operations), agent control and audit systems in closed factory networks and legacy environments (L6 governance and security), and converting agent performance into ROI and financial verification (L7 performance management).

The event begins with a presentation by Bespin Global CAIO Senior Vice President Seon-ho Han (한선호) on "AI transformation in manufacturing, a roadmap read through 7Layer." It will then feature Jae-yong Han (한재용), a manufacturing specialist at Amazon Web Services (AWS), on "successful AX transformation cases at global manufacturers," and Sang-won Hwang (황상원), an executive director at Bespin Global, on "AI agents for manufacturing sites, Bespin Global AX implementation cases." Hwang will introduce agent build experience and performance verification methods that moved beyond PoC into actual operations. Lastly, OpenAI account director Dong-jae Lee (이동재) will speak on "the meaning of ChatGPT Enterprise for manufacturing companies."

"The reason manufacturing AX stops at PoC is not technology, but because the structure that decides linkage to profit and loss and scaling is empty," Han said. "AX is not a problem of adoption but of alignment. The moment you define performance not by the number of PoCs but by work improvements verified by finance, the path to companywide rollout opens," he said.