Using AI in business process outsourcing.

 

One study examined the impact of GenAI on customer service agents at a United States-based business process outsourcing company, focused on the staggered deployment of a GPT-powered chat assistant firm serving SMEs, with some of the agents based in the Philippines and others in the United States and elsewhere. The study showed that AI significantly improved worker productivity across three key metrics, namely, reduced handling time per chat, increased chats handled per hour and successful chat resolution rates. Yet these benefits were not uniformly distributed; the most significant improvements were among less-skilled and newer agents while highly skilled and experienced workers showed minimal gains. This finding is particularly significant given the steep learning curve and initial lower productivity often associated with newer hires in the business process outsourcing sector. Interestingly, agents who adhered closely to AI recommendations demonstrated greater productivity gains, suggesting a link between AI engagement and learning. The agents sustained higher productivity even during software outages when AI assistance was unavailable, indicating a lasting impact on skill development. The study also considered the impact of AI on workers. Contact centre work often involves demanding overnight shifts and challenging interactions with customers, but the study showed that, when the workers were supported by AI, customers were impressed, less likely to question their competence and generally treated them better. This helped reduce employee attrition, particularly among newer hires. The researchers attributed these positive effects in part to the ability of the AI system to capture and disseminate best practices from high-performing agents. However, customer satisfaction can also be reduced if using AI makes interactions feel overly scripted and inauthentic. The study concluded that while AI assistance can enhance productivity and improve worker experience, it also creates incentives for firms to deskill positions and hire lower-skilled workers at lower wages. Companies could also eventually deploy even more advanced AI systems capable of entirely replacing human agents. While offering significant potential for companies, the long-term implications for workers remain uncertain and may depend on the strength of workers’ voices in workplace consultations or collective agreements. The findings are corroborated by another study involving 300 call-centre operators that showed that AI that automated repetitive tasks and provided real-time support could reduce stress levels among agents.