DX · AX
Digital Transformation / AI Transformation
What are DX and AX?
- DX = Changing how work gets done with digital technology
- AX = embedding AI into work and business
- AX comes after DX
What DX really changes is not systems but approval lines and work boundaries. If you have installed an ERP but approval requests still circulate on paper and sales data is locked in departmental Excel files, that is not transformation but expensive computerization. The same goes for AX. It is not about buying 100 generative AI accounts. It is about shifting 'who writes the first draft' from people to AI, and redefining the human role from author to verifier and decision-maker.
In practice, the conversation gets clearer once you separate three layers. Computerization (turning paper into files), digitalization (putting processes on a system) and transformation (changing the business model and organizational structure) are very different levels of difficulty. AX also gets harder in stages: repetitive document work, then internal knowledge search, then decision-making support, and finally building AI into the product itself. Automation such as RPA follows fixed rules, while AI answers by probability, so AX always keeps a human review step. One more point: AX does not work without DX. If three years of records exist only as scanned PDFs, there is no source data for AI to read.
The most common failure in the field is the state of 'adopted, but nobody uses it.' Run a single 2-hour company-wide lecture, hand out licenses, and three months later active users have dwindled to around ten. Rolling tools out before security guidelines are in place also leads to real incidents, such as client drawings or personal data being pasted into an external model. If you change only the tools and leave KPIs as they are, the time saved gets filled with other busywork, so results never show up in the numbers. Then at next year’s budget review, the conclusion hardens into "We tried AI and it doesn’t suit us." Once an organization fails this way, resistance doubles.
Seolyeong ParkGenerative AI and AX training expert with 17 years of hands-on experience and 200+ lectures, who gets non-specialists using AI right away
John KimJohn Kim, CEO of Nanum Consulting and author of “B2B Marketing Design”Digital transformation (DX) became a common term in the 2010s with the spread of cloud and mobile. AI transformation (AX) is the next-stage concept that emerged with the rise of generative AI (since 2022).
Consider a hypothetical case: Company A, an auto parts manufacturer with 320 employees and annual revenue of KRW 140 billion (≈ US$104,743,400). Producing a single quality-claim reply for a client took 6 hours on average, and the 9-person quality team handled 40+ cases a month, so overtime was routine. Here is the record of the 8 months that began in March 2024, when the CEO said, "Let’s try to cut this down with AI."
- The quality team lead broke the 6 hours spent on one reply letter into steps and timed each. 3 hours 40 minutes went to routine work: finding similar past claims and copying them into the template.
- An IT manager converted 1,860 claim documents from the past three years from scanned PDFs into searchable text. Their takeaway: "The cleanup took two months, longer than adding the AI."
- They ran a 6-week program for all 9 members of the quality team, with one 90-minute session a week. Instead of lectures, each week everyone brought a prompt for their own work and the team improved them together.
- After two cases in August where an AI draft got the root-cause analysis wrong, Article 4 of the quality manual was revised to require sign-off by a section chief or above on every final reply.
- As of November, the time per case fell from 6 hours to 2 hours 20 minutes, saving about 147 hours a month, and the quality team’s overtime pay dropped 38% year over year. Three departments, including Purchasing and Production Management, have since asked to adopt the same approach.