Before approving the next AI budget, check the network

Before approving the next AI budget, UK business leaders must ensure their networks can handle AI at scale once it moves beyond the pilot stage. Research shows that 55% of organizations plan to prioritize AI or ML investment over the next year, with one-sixth investing aggressively without sufficient evaluation. However, the ROI debate surrounding AI is too narrow.

While some benefits come quickly, others emerge from automation, productivity improvements, agility, resilience, faster decision-making, and new ways of working. Boards often fund AI as a transformation, then evaluate it like a short-term software project, which can lead to disappointment. When AI underperforms, the focus is often on the model, data, or the team responsible, but the real issue often lies in the network, security model, governance, and skills.

AI depends on a robust context, including the data, network, security model, governance, and skills. If these conditions are weak, even a promising use case can struggle to become durable and valuable in production. Among UK businesses where AI implementations failed to meet expectations, 24% cite inadequate network or connectivity performance as a contributing factor.

This highlights the importance of considering network performance when budgeting for AI projects. The network has become a critical part of the ROI calculation for AI initiatives, as it impacts data movement, workload routing, uptime, and cost accumulation. To ensure AI investment leads to durable value, boards should ask four questions before approving the next AI budget: where the data will move, whether the network can support AI beyond the pilot stage, what happens when something fails, and whether the organization has the conditions to scale AI.

These questions force boards to examine the infrastructure, skills, governance, security, data movement, and cost discipline required to support AI initiatives. While short-term measurement is important, it should not be used as an excuse to underinvest in capability building, which is essential for long-term value. As AI becomes increasingly integrated into operations, customers, fraud detection, forecasting, and field service, the network plays a crucial role in enabling AI-driven growth.