Google says companies need to understand AI token costs early so adoption does not turn into an unexpected bill surge. The warning comes as more businesses look to scale AI across their organizations.
Karim Siregar, Country Director of Google Cloud Indonesia, said companies should first calculate token costs and the business value they bring. He said those two points should be clear before any wider deployment begins.
“So we need to know how much the cost is, then whether the token consumption adds value to the business or not,” Karim said at a media briefing in Jakarta on Wednesday (15/7/2026).
Start with small pilots before scaling up
Karim explained that companies should master the cost and benefit calculation while they are still running small pilot projects. That way, they will not be caught off guard when AI implementation expands to the enterprise level and triggers a large jump in spending.
Google says it offers a complete solution to help companies estimate how many tokens are needed for specific AI capabilities. The goal is to let businesses forecast budgets more accurately from the start.
Better visibility and model selection
To help prevent unexpected token-cost spikes, Gemini Enterprise includes FinOps services that provide greater visibility into performance and cost. The feature is designed to make AI usage easier to track.
Moe Abdula, Vice President Customer Engineering Google Cloud Asia Pacific, said Google will also recommend the best model based on a company’s needs. Google currently offers more than hundreds of AI models.
“In addition, we have an intelligence layer, which we call the Vertex layer, that can understand the task being requested, to better suggest using this model because it is cheaper than other models,” Moe said at the same event.
Google says the approach helps companies choose models that match their needs while keeping token costs more efficient. The company wants businesses to scale AI without being surprised by expense increases.
FDE team expanded in Indonesia
Google is also expanding its Forward-Deployed Engineer (FDE) team in Indonesia. These are AI specialist engineers placed directly inside a company’s work environment.
Google says the team is meant to reduce implementation risks, speed up deployment, and maximize the value of customers’ Google Cloud investments. The support is intended to help companies make the most of AI innovation for business transformation through agentic AI.