Insights By Uday Birajdar, CEO & Co-founder, AutomationEdge
When discussing AI projects with CIOs, MSPs, and GSI leaders, a common concern that arises is the cost of tokens and ensuring that the budget for AI projects is not exceeded.
Earlier this year, our team encountered a situation at an insurer where token consumption exceeded estimates due to unoptimized prompts in historic policy documents. This resulted in the process being halted within a week. A similar issue occurred in an underwriting workload.
It’s important to note that such challenges are not limited to the service desk workload but can impact various areas within organizations. Estimating consumption is often a trial-and-error process until it reaches production, leading to unexpected costs.
In a surprising revelation, Uber’s CTO mentioned that the company’s entire AI budget for 2026 was depleted within just four months.
These instances serve as a warning for service providers as the same challenges can affect the service desk operations, making them more vulnerable to budget overruns compared to larger companies like Uber.
