Steve Barrett

As AI moves from experimentation to production, enterprises are faced with an uncomfortable reality: costs can escalate faster than they expected. In fact, many well-known companies are reportedly tightening controls on AI usage. They are doing this not to slow innovation, but to ensure their investments remain sustainable. The problem is that, unlike traditional software, AI is priced by consumption. Every prompt, model invocation and inference request carries a cost.

Last year, many UK businesses were affected by one of the largest IT outages in history. Systems inexplicably failed due to a growing dependency on third-party solutions. How do they maintain operational resilience in the face of mounting risks—especially in cloud environments, where services are distributed and deeply interconnected? In this article, Steve Barrett, the VP of EMEA at Datadog, explores how organisations can implement risk management procedures to achieve incident readiness.