The rising costs of enterprise AI, particularly due to token-based pricing models, are causing concerns among CFOs as organizations struggle to manage expenses and tie costs to value. Celonis' Field CTO, Manuel Haug, outlines a three-stage adoption process for agentic AI and emphasizes the need for a shared architectural context to optimize costs and improve decision-making in enterprises.
For professionals tracking enterprise AI and SaaS, the key insight is the potential transition from seat-based to consumption-based pricing models for agentic AI, akin to infrastructure pricing on platforms like AWS. This shift necessitates strategic planning to manage costs effectively, as token utilization can rapidly escalate without established governance frameworks, highlighting the importance of aligning AI investments with business outcomes to prevent budget overruns.