The AI pricing conundrum – it started as a nightmare, now it's worse (www.computerworld.com)

The ongoing struggle with AI pricing models has become more complex as enterprise IT leaders grapple with aligning costs to delivered ROI. Traditionally, IT executives have found it challenging to establish appropriate payment structures, often feeling disconnected from the actual use cases as line-of-business workers experiment with AI technology independently. Current pricing models, such as paying per token or per task, fail to reflect the true value delivered by AI systems, which complicates budgeting and ROI justification. Experts, including Irfan Khan from SAP and Justin Greis from Acceligence, highlight that the disconnect stems from two contrasting priorities: enterprises seek value-driven pricing while AI vendors prefer predictable consumption-based models. The implications of this pricing dilemma are profound, as enterprises could face operational risks if AI systems are primarily incentivized for measurable outcomes without proper oversight. As Greis warns, metrics can drive AI behavior in ways that may be harmful if the focus shifts solely to efficiency or cost-cutting without considering ethical or strategic alignments. To counter these risks, establishing an AI oversight committee could ensure projects are comprehensively evaluated, fostering accountability among executives and aligning AI objectives with acceptable outcomes. This approach may ultimately provide a way for enterprises to advocate for fairer pricing models that reflect both the value and risks associated with deploying AI technologies.