Korean Survey on AI Customer Service: Expected Job Cuts Did Not Materialize; Clear Human-AI Handoff Standards Prove Key to Success
A survey in South Korea reveals that more companies hired staff than cut them after adopting AI for customer service, with predefined human-AI collaboration standards significantly boosting success rates.
South Korean HR tech company Remember&Company released its 'AI Customer Experience Report' on October 30, detailing the actual workforce changes and operational outcomes for major Korean enterprises after implementing artificial intelligence in customer service. The report is based on a survey of CS practitioners and decision-makers at key Korean firms.
The data shows that while 36% of companies without AI expected headcount reductions upon adoption, only 18% of those actually operating AI reduced their staff. Conversely, 29% increased their customer service personnel, and 45% opted to reallocate tasks rather than cut jobs. This indicates that the widespread expectation of 'AI inevitably leading to customer service layoffs' has not fully materialized on the ground.
The report also highlights that customers still prefer human interaction even after AI implementation. 50% of companies reported that customers most frequently requested transfers to human agents, and 45% faced cases where customers resubmitted the same inquiries. Although response speeds improved, they did not meet customer expectations for resolution. Furthermore, there is a gap between internal metrics (such as response time and processing cost), which showed a 65% improvement rate, and external customer perception metrics like satisfaction, which showed only a 52% improvement rate.
A critical factor determining the success of AI customer service projects is not just AI performance, but whether clear 'role division standards between AI and human agents' were established beforehand. Among companies whose AI results exceeded expectations, 78% had such standards in place, compared to only 32% among those who fell short. However, currently, only 8% of AI-adopting companies have documented these standards, with most relying on employee experience.