As companies expand their use of artificial intelligence, the ability to apply AI to actual work and verify the results is becoming more important than simply building systems. VNTG, an information technology services firm specializing in manufacturing, analyzes clients' work environments to identify where AI can be applied and supports the entire process from planning through development and operations, helping companies with what it calls AI transformation, or AX.
VNTG provides IT services that support corporate digital transformation and AX, drawing on its experience building and operating systems at manufacturing sites. Rather than stopping at developing the functions a client requests, the company focuses on finding problems that need fixing in day-to-day operations and proposing technical solutions.
VNTG first defines the problem that needs solving at a client's worksite, then quickly tests whether AI can be applied and what effect it would have. It analyzes workflows to find tasks where improvement is likely to be greatest, runs an initial verification and then decides the actual scope of development. The distinguishing feature is that it confirms the effect of applying the technology first and shapes the project around those findings, rather than fixing the development scope in advance.
That differs from how the IT services industry has typically operated. Under the conventional approach, a client finalized the development scope and requirements, signed a contract and then began building the system. But as AI technology advances rapidly, that approach is making it harder to respond to companies' shifting needs. Contract talks and preparation can take months, during which the technology a company planned to adopt or the problem it set out to solve may change.
"Even when we want to define a problem based on what the worksite needs and apply an AI solution quickly, the preliminary contract talks and procedures alone take months," said Lee Seung-eun, vice president of VNTG. "While the contract is being negotiated, the original task often loses its timeliness or turns into an entirely different problem."
The company also judges the results of an AI rollout not by whether a function works but by quantitative measures such as reduced working hours and improved productivity. "Cutting a task that took 10 hours down to one hour, or having two people finish analysis and documentation work that three people handled, in half the time — that is intuitive value," the vice president said. "The burden of searching for and analyzing information falls, while the quality and volume of the insights produced rise."
One approach drawing attention in this context is the forward-deployed engineer, or FDE, model. Under FDE, technical staff are placed directly at a client's site to identify problems in its operations and to develop and apply the AI solutions needed. Where consulting firms previously proposed directions for improving operations and IT services firms carried them out, FDE ties problem diagnosis and technology development closely together.
VNTG is developing this site-centered approach to problem-solving into an integrated service system that links planning with system construction and operations. The company's view is that the results it initially expects can be achieved only if it also manages the process of applying an AI system to actual work and running it reliably after the system is built.
To support that, VNTG is also strengthening the capabilities of its own staff. The company is reworking how it organizes its operations so employees can understand clients' business problems and propose technical solutions. "The limits of one person's experience and knowledge must not become a bottleneck for a client's business," the vice president said. "We need to build a system in which the company's entire body of expertise and personnel comes together around the client's problem."