OpenAI is targeting the business and knowledge work market with its ChatGPT Work, which is based on the coding agent Codex.
On August 20, OpenAI held a demonstration of ChatGPT Work in Gangnam, Seoul, sharing real-world use cases from local experts. ChatGPT Work is designed to perform various tasks such as data research, analysis, and document creation, leveraging the capabilities of OpenAI's Codex coding agent.
Eum Eun-sook, head of communications at OpenAI Korea, highlighted the competitive edge of the coding agent's capabilities. She stated, "Differences in coding abilities lead to variations in results among models. At this point, I believe OpenAI has a competitive advantage in this area."
The number of users is rapidly increasing. According to OpenAI, Codex surpassed 3 million weekly active users (WAU) globally in April. Since the launch of ChatGPT Work in July, the combined WAU for both products has exceeded 10 million. The agent-based AI, which initially focused on development tasks, is now expanding into various knowledge work areas, including research, analysis, and document creation.
As a result, domestic companies and institutions are increasingly adopting agent-based AI. OpenAI reports that various organizations, including Samsung Electronics, Samsung Display, and Seoul National University, are utilizing related services.
In the corporate sector, the value of AI is growing as it evolves from simple question-answering chatbots to agents capable of performing actual tasks. Eum noted, "Previously, companies would ask, 'What is this?' and receive answers, but now agents are emerging that can complete actual work, increasing their utility."
The expansion of ChatGPT Work is expected to contribute to OpenAI's revenue growth. Eum stated, "It seems likely to be very helpful for revenue." However, she clarified that there are currently no specific plans for an initial public offering (IPO).
Meanwhile, on the same day, Sim Dae-yeol, an engineer at OpenAI Korea, demonstrated how to use ChatGPT Work to gather necessary information from work tools like Slack and Google Drive to create presentations, and how to regularly check and update project information through scheduled tasks.
Real-world use cases in education and research were also shared. Go Gil-gon, head of the Informationization Office at Seoul National University and a professor at the Graduate School of Public Administration, introduced a research case where Codex was used to collect and structure statistics, laws, policy documents, and academic materials by country, automating everything from data analysis to drafting reports.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.