Corporate priorities around artificial intelligence (AI) have shifted from "how much AI to deploy" to "how safely and measurably AI can deliver results." The change reflects how AI has become part of daily business operations, from leveraging large language models (LLMs) such as ChatGPT and Gemini to building AI agents tailored to specific work environments.
According to a March survey of 600 chief information officers (CIOs) worldwide by global AI platform company Dataiku, 95% of respondents said they report AI performance to their boards. Yet only 25% of CIOs said they could monitor all AI agents created within their work environments. As the gap between AI adoption speed and management capability widens, the importance of "AI governance" has grown.
Dataiku APAC VP: "Governance Essential to Unlock AI's Potential"
Andrew Boyd, senior vice president for Asia-Pacific and Japan (APJ) at Dataiku, who recently met with The Seoul Economic Daily, offered the same diagnosis. "For AI orchestration—connecting enterprise data with AI models—and talent to fully realize their potential, governance must come along," Boyd said. "Access must be limited to permitted data, companies must be able to explain how the AI agents used in practice were developed, and their outcomes must be measurable in a visible way."
Founded in France in 2013, Dataiku is an AI and machine learning company. It helps enterprises safely use the latest technologies, such as LLMs and AI agent creation, while maintaining their existing IT infrastructure. A distinctive feature of Dataiku's AI platform is that it allows immediate use of corporate data scattered across on-premises servers and the cloud without needing to move it. About 25% (125 companies) of the top 500 firms on the Forbes Global 2000 are Dataiku customers. The company began operations in Korea in 2021 and counts LG Chem and Doosan Enerbility among its clients.
"Korean CEOs Feel Job Threat From AI Failure—Twice the Global Average"
Before joining Dataiku in March, Boyd worked at major Big Tech firms including Microsoft (MS) and Amazon Web Services (AWS), supporting various Korean companies in their global expansion. Familiar with the Korean market, he said, "Korean corporate CEOs are highly receptive to AI but also face strong demands to comply with regulations. In our survey, 20% of Korean CEOs answered that 'their jobs would be threatened if AI strategy and execution went wrong,' which is far higher than the global average of 10% and the U.S. figure of 4%."
He added, "Aggressively pursuing innovation while managing risk well is a defining characteristic of the Korean economy. Given that Dataiku has particular strengths in governance, Korea is a very important market for us."
Dataiku, a 'Governance-Optimized' Platform That Also Measures Agent ROI
Dataiku's new brand "Platform for AI Success," unveiled in March, includes a wide range of solutions designed to deliver governance. A representative example is "Agent Management." This solution specializes in managing AI agents created by employees during corporate workflows and measuring return on investment (ROI).
"We are currently providing this solution as a preview to select customers, including in the Asia-Pacific region," Boyd said. "One large enterprise customer believed its employees were creating about 500 agents per week, but when checked through Agent Management, 2,000 agents had actually been generated." While AI was widely used across operations, agent management had not been thorough.
"Agent Management can solve these issues," Boyd explained. "It allows companies to track and manage whether agents access only permitted data, whether they operate as intended, and whether they are delivering ROI." Dataiku plans to officially launch Agent Management soon.
In addition, Dataiku plans to roll out "Reasoning Systems" by industry, which understand an organization's business context and enable AI to make decisions. It also offers "Cobuild," which allows non-developers to build AI agents using natural language.
Results at LG Chem and Doosan Enerbility… "Will Help Spread Korean Innovation"
Dataiku plans to pursue the Korean market more aggressively going forward, building on confidence gained through collaborations with clients. LG Chem built a manufacturing quality control system based on Dataiku's platform and achieved early detection of anomalies in manufacturing processes. As a result, LG Chem's U.S. and European bases have also begun using the system.
Doosan Enerbility applied AI to predict molten steel output from electric furnaces using Dataiku's machine learning operations (MLOps) platform. Recognized for raising forecast accuracy to the 98% level, the project won at the 2023 IDC Future Enterprise Awards. "Dataiku has been steadily growing its business in Korea, posting double-digit revenue growth," Boyd said. "I will continue to spend a lot of time in Seoul and contribute to spreading Korean innovation to the rest of the world."
