Could AI agents be causing surprise bills at your business? You're not alone (www.techradar.com)

A recent report by Splunk reveals that businesses are grappling with unpredictable AI costs as adoption of agentic AI increases. Approximately 60% of firms report that spikes in AI bills create greater financial strain than a higher, consistent baseline. This shift to consumption-based pricing models complicates cost management, with over half of the organizations unable to effectively control their AI expenses. Compounding this issue, AI agents can independently vary their token consumption based on the complexity of tasks, leading to financial unpredictability and heightened worries about telemetry costs exceeding infrastructure expenses. Despite these challenges, the rise of AI has fostered greater collaboration between observability and security teams, with 81% of respondents noting improved cooperation in monitoring AI workloads. This convergence is crucial, as 91% of observability leaders can now accurately attribute costs to specific models and agents, significantly aiding in understanding AI's return on investment. However, the preference for flatter, predictable pricing models remains, with three in five organizations indicating that unexpected billing spikes are more burdensome than a consistently high cost, underscoring the need for businesses to rethink their strategies around AI cost management.