Financial operations (FinOps) have evolved from a reactive method of monitoring and managing rising cloud costs into a proactive AI-driven operating model spanning engineering, finance, and transformation production. “FinOps is now facing its greatest challenge to date,” says…
arlotte Dunlap, Research Director for GlobalData. “It’s tasked with managing AI spend, specifically the fast-growing cost of tokens associated with the use of LLMs [large language models]. With new reasoning-based models and resulting token requests comes the need to figure out token efficiencies for gaining necessary value to justify costs, as just one of the challenges practitioners face.” The increased multi-step reasoning characteristics associated with models mean the number of tokens per request is increasing substantially
More tokens mean longer sequences are fed through large models. Considering the complexity of the unit economics of AI features, also known as ‘tokenomics,’ it’s critical to have FinOps practices in place at the application and workflow layer for mapping AI costs. Another growing trend in FinOps is its application to money movement within the financial sector, moving from reactive to proactive financial oversight.
AI-driven tools can interpret signals in payments, accounts, and risk in real time, so organisations can anticipate and adapt to change proactively. Going beyond just automation, AI can enable predictive controls, exception handling, and commercial insight, helping to ensure meaningful decision-making. Increased global regulations are a driving force behind a more modern FinOps, which emphasises building standardised, configurable, and repeatable processes for regulatory controls and reconciliation.