In the Era of AI Transformation 2.0, AI Becomes a Colleague, Not Just a Tool

[Financial News] AI tools have made individuals faster at their work, but the workflows through which organizations make decisions have not changed. This is the context behind a report by NANDA, affiliated with MIT, which found that 95% of companies’ generative AI pilots failed to produce meaningful financial results. Adoption itself is already widespread: one survey found that 69.3% of companies in South Korea are planning or pursuing AI transformation (AX). The problem is not adoption, but what comes next.
Companies have faced two major obstacles in their efforts to achieve AX. The first is the trap of partial optimization: assuming that adopting tools is itself AX. Tools for tasks such as summarization and automatic classification may reduce the time spent on a particular step, but the overall workflow still takes just as long. Because AI tools are used separately for different purposes, a new “bottleneck” has emerged: people must manually process and review the output of one step before it can be used in the next. The second obstacle is data fragmentation. Companies have plenty of data, but it is scattered across individuals’ computers and departmental folders. Since AI cannot learn from or make use of an organization’s historical data and context, it produces only one-off answers and fails to provide insights deep enough to inform real decisions.
A different perspective is now taking hold, particularly among companies quick to adopt new technologies. Rather than viewing AI as a simple “tool” for specific tasks, they see it as a “capable colleague” working alongside them across the entire workflow. Vertical AI, in particular, understands context deeply and, within a specific domain, goes beyond handling routine tasks to expertly complete the entire workflow, from work design to analysis and reporting.
In South Korea, this shift is bringing tangible change to “consumer data collection and analysis,” an area that has long been plagued by decision-making bottlenecks. The greater the data gap between departments—or the more an organization lacks a dedicated data team—the greater the impact. OPENSURVEY INCORPORATED’s DataSpace, for example, uses AI as an expert collaborator throughout the entire process, from survey design to data cleaning and analysis to report generation. It reads documents and data scattered across the organization and understands the context of past research, taking on the role of a “capable expert” who offers in-depth insights. As a result, research workflows that previously took at least two weeks are dramatically compressed to around a day. Without hiring additional specialists, existing staff can increase their output and accelerate decision-making, translating directly into clear ROI (return on investment).
Companies now need to change the question they ask about AX. It is not “What tools should we buy?” but “Which colleagues should we work with?” Simply adding more tools cannot resolve fragmented ways of working or decision-making bottlenecks. What matters is having AI participate throughout the work process to shorten and redesign the workflow itself. If AX 1.0 was a competition over individual tools, workflow is the competitive battleground in AX 2.0. The gap will widen further between companies that have adopted AI tools and those that bring AI in as a colleague and transform their workflows.
 
Hwang Hee-young, CEO of OPENSURVEY INCORPORATED