"Light bulbs made factories brighter, but the assembly line changed industry. The value of artificial intelligence (AI) does not end with automating a single task. The real value lies in completely redesigning how corporations work, their decision-making structures, and how systems achieve consolidation."
Lee Su-jeong, president of IBM Korea, said this at the IBM AI Summit Korea 2026 held on the 1st at the Westin Seoul Parnas in Samseong-dong, Seoul. Lee said AI is the key driver that innovates the entire operation of corporations, emphasizing that in the AI era, corporate competitiveness is divided not by "better models," but by the "capability to achieve consolidation of AI into operations."
Lee said many corporations have built AI models and data capabilities but are struggling at the stage of integrating AI into core processes and linking it to outcomes, and introduced Bob, an AI development platform IBM recently released. Unlike existing AI tools that stop at code writing, Bob is an agent-style platform that automates the software development life cycle (SDLC), including development, testing, deployment, and modernization. Depending on the nature of the task, it automatically distributes work across multiple models such as Anthropic Claude, Mistral, and IBM Granite.
Lee said, "The moment an agent carries out real work, it is very important to design which data it accesses and what it executes under what permissions," adding, "Through this process, AI becomes not just a development tool but the operating system that supports the entirety of corporations."
In the keynote that followed, Neil Sundaresan, senior vice president overseeing IBM Automation & AI, said, "Across the product development life cycle (PDLC) or SDLC, Bob acts as a companion at each stage," adding, "More than 160,000 users inside IBM are using Bob for their respective tasks." For example, assuming a task that interprets data coming from multiple sources and converts it into an API (application programming interface), in the past developers had to write conversion rules themselves, which was labor-intensive, but Bob builds APIs with natural-language instructions instead of developers writing rules one by one. It goes further by filtering unnecessary data, classifying patterns, devising an execution plan on its own, and then generating not only the actual runtime code but also the test code that validates it and the explanatory documentation, all at once.
Sundaresan said, "Over the next five years, 1 billion applications (apps) will be developed," adding, "Agents that operate across individual apps are likely to reach 2 billion, at least twice as many as the apps."
At the event, Kwon Hae-seong, vice president in charge of AT/DT platform development at SK Telecom, said SK Telecom is overhauling its business support system (BSS) to fit the AI era. BSS is the system responsible for charging and billing, customer management, and product and rate plan management. Kwon said, "We are not layering AI on top of the existing system; we decided to change the system itself so AI can read and act," explaining that after about three years of review, the company decided on a full BSS overhaul. Replacing a major carrier's BSS even once takes more than five years and astronomical expense, but the company ultimately decided on a full overhaul, judging that in the AI era this is unavoidable.
Park Jae-man, Deputy Minister of AI and Digital at KB Securities, introduced the case of building Kkaebi, an AI service for employees, and said, "The starting point of successful AI is not the model but the data." Park said, "The entity that now brings data to end users is not a person but a large language model (LLM)," explaining that the team focused on enabling the LLM to find data quickly and understand it accurately. Park added that beyond using Kkaebi for internal task support and customer-facing asset management, the company will build separate AI services based on it and pursue a business-to-business (B2B) business.