Johannesburg: Artificial Intelligence (AI) is making significant inroads into the Business Process Outsourcing (BPO) industry, promising enhanced efficiency and improved customer engagement. However, contrary to expectations, AI is increasing operational costs within South African call centres.
According to African Press Organization, as AI tools become integral to customer service environments, operators are realizing that the cost of implementing AI extends beyond software licenses to the infrastructure required to support it. The demands for processing power have surged, necessitating significant investments in computing infrastructure to handle AI-driven applications.
The backend impact of AI adoption has been underestimated by many providers. AI requires substantial compute power, memory, and low-latency environments, leading to increased infrastructure expenses. BPOs are now faced with the challenge of either upgrading endpoint devices to handle AI workloads locally or shifting the processing burden to backend servers, both of which come with significant financial implications.
A noticeable shift in hardware requirements is occurring, with many providers transitioning from traditional Intel i5 deployments to more advanced i7 devices to support AI-enhanced workloads. Alternatively, some operators are opting to centralize AI processing on servers, avoiding desktop upgrades but necessitating robust server infrastructures with higher compute density and advanced networking capabilities.
The financial strain is evident as the cost of expanding server environments to accommodate AI workloads rises. This is compounded by the global demand for AI-capable hardware, driving up prices further.
To mitigate these costs, some organizations are exploring off-premises solutions, utilizing hyperscale providers like Amazon Web Services or colocation environments. While this model alleviates upfront infrastructure investment, it introduces ongoing rental and operational costs that require careful management.
AI is fundamentally altering the economics of the BPO industry, shifting the focus from labour efficiency to infrastructure efficiency. The procurement model is evolving, with many operators moving away from capital expenditure projects towards leasing and rental models to deploy backend AI infrastructure.
The shift towards operational expenditure models allows for cost distribution over time, maintaining flexibility as AI requirements evolve. This reduces the risk of overinvesting in hardware that may quickly become obsolete.
The reality is that AI is not inherently reducing operational costs within BPOs; rather, it is shifting costs from labour to infrastructure. The emerging competitive advantage in the BPO industry may soon revolve around the ability to power AI at scale, challenging traditional cost models and reshaping the industry landscape.