Companies struggle to convert AI investment into real business results

Frontline staff to design and build their own AI-powered tools

AI applied to tire pattern generation and demand forecasting

'Defining the problem matters more than mastering the technology'

Kim Sung-jin, executive vice president and chief digital and information officer at Hankook &, delivers a keynote address at the industrial AI conference Attention 2026 at the Westin Josun Seoul in Gangnam-gu, Seoul, on Thursday. [Hankook & Group]
Kim Sung-jin, executive vice president and chief digital and information officer at Hankook &, delivers a keynote address at the industrial AI conference Attention 2026 at the Westin Josun Seoul in Gangnam-gu, Seoul, on Thursday. [Hankook & Group]

Companies are racing to adopt generative AI, yet fewer than expected are translating that investment into tangible business results such as higher sales or productivity gains, according to an assessment shared Thursday. Hankook & Group says it plans to go beyond simply adding AI tools — instead, it wants frontline employees to redesign their own workflows and build the software they need themselves.

Hankook & Group outlined its AI transformation strategy and early use cases Thursday at Attention 2026, an industrial AI conference held at the Westin Josun Seoul in Gangnam-gu, Seoul.

Attention, an annual conference organized by MakinaRocks since 2024, ran this year under the theme "AI Lands on the Ground." The event drew about 1,600 pre-registrations, primarily from manufacturing, energy, semiconductor, defense and public-sector organizations, with more than 800 people attending in person.

Kim Sung-jin, Executive Vice President of Hankook &'s Digital Strategy Division and the group's chief digital officer and chief information officer, delivered the grand keynote on how AI is transforming corporate operations. Kim oversees IT, digital transformation and AI transformation for the group.

Kim opened by diagnosing what he called an "ROI plateau" — a pattern in which companies that have adopted generative AI are failing to see the returns they anticipated.

Fewer than 30 percent of companies that have introduced AI have managed to generate real business value from it, with many still stuck at the proof-of-concept stage, he said. Fragmented data spread across research and development, production, quality control, logistics and finance — combined with legacy IT systems that take a long time to move from planning to deployment — are the main obstacles, Kim explained.

Hankook &'s answer is what it calls "AI-driven business process innovation." Rather than layering AI features onto existing workflows, the group intends to redesign those workflows from the ground up with AI use built in from the start.

The foundation is data integration. The group is building a cloud-based data platform to connect information that has been scattered across its businesses. It is also establishing a single source of truth — a group-wide standard dataset known internally as SSOT — to reduce inconsistencies that arise when different departments interpret the same data differently.

The way employees develop software is also changing. Hankook & Group is rolling out what it calls "AI vibe coding" across its operations — a practice in which frontline staff build the tools they need by conversing directly with generative AI.

In the past, when employees needed a new program, they would compile requirements and hand them off to a development team, then wait through a build-and-validation cycle. With AI, the people who know the work best can define the features they need and create simple programs or task tools themselves, Kim said.

AI has also entered the tire development process. Hankook Tire has partnered with MakinaRocks to build a tire pattern generation AI that designers use to produce a wide range of tread patterns. The system generates multiple design options, which engineers then evaluate for performance and aesthetics before feeding the results back into the development cycle.

The group has applied AI to demand forecasting as well, building a proprietary model that analyzes 17 internal and external variables to project future sales volumes and support supply chain decisions on production and inventory.

Kim said the ability to identify which problems need solving will matter more than technical proficiency with AI itself. Even as AI lowers the barrier to software development, defining the problem and judging whether AI-generated outputs actually help the business remain fundamentally human responsibilities, he said.

Hankook & Group has also been expanding AI use within its internal work environment. On Tuesday, it relaunched Arena, its group-wide work platform, with a major overhaul that added capabilities for automating repetitive tasks and creating personalized work bots for individual employees, on top of existing document search and summarization features.

"Enterprise AI innovation going forward will not be the exclusive domain of technology specialists — it will shift toward frontline practitioners who understand the business most deeply and think with the greatest clarity," Kim said. "Combining data, AI and on-the-ground expertise to continuously generate real business results will become the defining competitive advantage for companies."


kwater@heraldcorp.com