[Edaily KIM EUNG-TAE Reporter Ahn So-hyun] #Public relations agency A has recently been requiring its employees to take training in “vibe coding” (a method in which a person describes the required functions in natural language, and generative AI writes the code to develop the program). As more clients are using AI to create simple promotional materials in-house, the agency aims to secure new business by developing programs capable of performing complex tasks. Industry observers interpret this as an effort to streamline employees’ workloads using AI—rather than hiring new staff—amid deteriorating business conditions.
Mid-sized and small businesses are rapidly bringing in-house the operational systems they have traditionally outsourced by leveraging AI. This is interpreted as a strategy to reduce costs and improve processing speed by delegating tasks to AI solutions. As a result, outsourcing firms are facing a shortage of work, while AI startups that support workflow simplification are on the rise.

◇ K-Beauty Begins in-Depth Corporate Transformation with In-House AI
According to sources in the mid-sized and small business sector on the 6th, the cosmetics Original Design Manufacturer (ODM) firm COSMAX, INC. is expanding the in-house development of its core business systems using AI. The company plans to carry out 27 projects across seven areas—including key functions such as supply chain, cost management, and human resources, as well as logistics, production, data, security, and infrastructure—by next year using its own in-house staff, without relying on outside contractors.
To carry out these projects, COSMAX, INC. has hired a large number of AI developers this year. The company believes that as AI accelerates development speed, the number of tasks that can be performed in-house will increase, and the demand for development personnel capable of handling everything from design to verification and operation will also grow.
COSMAX, INC. has also made “AI-native development” a core strategy in its business plan for next year. The company plans to establish a company-wide system where in-house developers, who understand on-site operations, use AI to directly build the systems necessary for research and production and can quickly improve those systems as customer needs change. The ultimate goal is to reduce the time required for information sharing and decision-making across departments—from product development to production—and to develop and supply products that meet customer needs more quickly. COSMAX, INC. stated, “By shifting development and maintenance—which were previously outsourced—in-house, we will link AI-driven productivity improvements to reduced reliance on outsourcing and the securing of internal expertise,” adding, “We expect to save billions of won annually in external service costs by resolving key challenges on our own.”
Another cosmetics ODM company, KOLMAR KOREA Group, has also begun actively implementing AI projects to achieve its goals of improving operational efficiency and reducing costs. Last year, KOLMAR KOREA Group established the “AI Challenge Lab” and is currently carrying out a project to roll out AI utilization strategies across the entire company. According to the company, it has carried out over 50 AI projects to date, resulting in annual cost savings of over several billion won and improved operational efficiency. Furthermore, the company held an internal AI contest and received over 200 proposals for AI applications; it has selected key projects for a fast-track program and is working with practitioners and developers to further refine them.
Kolmar Group plans to launch an “AI Portal” within the year to accumulate and share AI assets across its affiliates, and to actively recruit AI talent. KOLMAR KOREA official stated, “We are expanding the application of AI beyond specific organizations to incorporate it into the actual operations of each department, including R&D, sales, and production,” adding, “While replacing existing outsourced and licensed services with our in-house AI, we are simultaneously hiring specialized personnel in areas such as AI operations, infrastructure, security, governance, and monitoring, and we are continuing to strengthen AI training for our employees.”
◇As AI Becomes Embedded… Outsourcing Firms Struggle, Startups Rise
As major mid-sized and small-to-medium-sized enterprises (SMEs) scale back outsourcing in favor of AI, outsourcing firms are facing a crisis of dwindling workloads. The industry consensus is that offers for simple, one-off tasks—such as research, advertising, and content production—that can be easily handled by AI have virtually disappeared. An employee at a public relations agency said, “Recently, as clients have adopted generative AI, such as ChatGPT, company-wide, the workload for agencies has decreased significantly,” adding, “In the past, we would write press releases directly and bill clients for them, but now clients often use AI to write them on their own.” The employee continued, “The areas that can be replaced by AI will expand faster than expected.”
The crisis facing the outsourcing and agency industries is a global trend. In a survey conducted this year by the U.S. advertising software company “Basis” of approximately 200 U.S. advertising agency professionals, 65.3% of respondents reported that, within the past year, clients had shifted some of their existing agency work in-house. Additionally, about 90% of respondents said that AI poses a threat to agencies’ primary revenue streams.
In contrast, startups providing specialized AI solutions to streamline operations are growing rapidly. Global beauty company APR(278470)has adopted a legal management AI solution from the startup “Bapteul” that automatically analyzes contract clauses and detects legal risks to handle related tasks. It has also adopted an AI monitoring solution from the startup “MarkVision” to address issues with counterfeit cosmetics.
Small and medium-sized enterprises (SMEs) are also actively adopting AI customer service solutions to improve operational efficiency. According to the startup Channel Corporation, the number of channels opened for its AI customer service solution “ChannelTalk” grew fivefold in five years—from approximately 50,000 in 2021 to over 240,000 as of September 2026.
An official from an agency in the industry stated, “We must take note of the fact that we’ve reached a point where any company can freely use AI tools,” adding, “Agencies will need to shift their role to focus on strategies and specialized tasks that are difficult to replace with AI.”