AI will reshape more jobs than it eliminates, concludes the BCG Henderson Institute. Within two to three years, 50 to 55 percent of all jobs in the US will change drastically in nature. Complete job replacement will occur within a timeframe of four to five years for 10 to 15 percent of jobs, including customer contact jobs.

 

The program BCG research distinguishes six categories of AI impact on functions. Two factors are decisive in this regard: the extent to which AI replaces human work (substitution) or enhances it (augmentation), and the extent to which productivity gains contribute to growth.

Six categories of impact

The BCG model divides functions – and the impact of AI on jobs – into six segments. 'Amplified roles' (5 percent of jobs) are enhanced by AI and grow along with increasing demand, such as for software engineers. 'Rebalanced roles' (14 percent) are redesigned: routine tasks disappear, while more complex responsibilities increase. 'Divergent roles' (12 percent) present a mixed picture: entry-level positions disappear, but senior positions remain or grow.

'Enabled roles' (23 percent) receive AI as a daily tool without the job fundamentally changing. And for 'limited-exposure roles' (34 percent), the impact remains limited for the time being. 'Substituted roles' (12 percent) are hit the hardest, with net job losses. According to BCG, contact center jobs belong to this last category.

In these substituted positions, AI can directly replace human employees in terms of core tasks. Efficiency gains in this function lead to net job losses. BCG also states that downward wage pressure arises on the remaining positions.

Contact centers as an example of substitution

BCG explicitly cites the call center employee as an example of a position that falls into the category of 'substituted roles'. According to BCG, the work of a customer contact employee consists largely of structured interactions: looking up account details, checking history, explaining policies, and scripted troubleshooting. As soon as AI systems can reliably handle these routine questions end-to-end, fewer employees will be needed. Moreover, the demand for customer service is limited by the size of the customer base and the frequency of contact moments. Therefore, productivity gains here do not lead to more work, but to fewer jobs, according to BCG.

At the same time, BCG notes that AI in the contact center is among the most mature AI applications, but that its market penetration remains limited. In practice, full substitution proceeds more slowly than augmentation because it requires processes to be thoroughly redesigned and implicit knowledge formalized. Moreover, employees often form an indispensable escalation layer, according to BCG.

The researchers state that in some cases, employees can advance to positions with more added value, focused on relationship management and proactive risk management. However, the total number of jobs in the role of call center agent will decrease as the most structured tasks are taken over by systems.

Consequences for the labor market

BCG warns of four side effects of the rise of AI. First, scaling up and relocating employees requires a massive investment in reskilling. Second, the number of entry-level positions is shrinking, because it is precisely the structured routine work performed by junior employees that is being automated first. Third, qualification requirements are rising: the remaining positions demand more expertise, judgment, and responsibility. And fourth, cognitive load is increasing now that repetitive tasks are disappearing and primarily complex thinking remains.

Method – The researchers analyzed the impact of AI on employment in three steps. First, the automation potential at the task level was determined for each role, based on detailed occupational data from sources including Revelio Labs and O*NET. Tasks are classified as automatable if they do not require physical presence, emotional intelligence, or open reasoning, and can be performed based on rules. The threshold was set at 40%. Next, it was determined whether AI replaces (substitution) or enhances (augmentation) human labor, based on two dimensions: the degree of human interaction and judgment, and the structure and repeatability of the work process. Finally, the scalability of demand was examined: do productivity gains lead to increased output and thus more work? This was assessed via price elasticity, open vacancies, and structural scalability. The model was applied to approximately 1.500 roles and 165 million jobs in the US and focuses on existing AI capabilities, particularly LLMs.

(BCG)

Featured, Human Resources, Research