There was a time when companies roughly knew what software would cost for the year. They would purchase a license and assign it to each employee. AI has changed this dynamic. As AI providers and coding tools move beyond fixed subscriptions toward token-based and other usage-based charges, giving an employee access to a model can also mean signing the company up for an open-ended spending commitment.
The need to understand AI spend and then control it within a large business depending on multiple AI providers, has opened a new category of tooling. Wiv, which helps enterprises and managed service providers manage their cloud and AI costs, has launched Wiv Air to keep AI usage inside defined budgets. The product connects to providers and tools including OpenAI, Anthropic, and Cursor, so companies can allocate budgets, enforce spending limits, and adjust access for every department, team, and employee. Teams keep using the AI tools they already rely on while the business retains control of the spend.
A common approach to AI cost visibility places a gateway or proxy in the path of every request. Wiv Air takes a different route, connecting directly to providers through their APIs, so it does not reroute AI traffic, require installed software agents, or add latency to requests. It syncs with identity providers such as Microsoft Entra ID to attribute usage to individual employees, teams, and departments. Finance and business leaders then share a single view of usage, costs, and remaining budgets, with spending from the company’s own AI products shown alongside employee usage.
Most finance teams are struggling to keep pace with the speed of AI adoption. In a survey of nearly 400 businesses, only 11% could accurately forecast their AI spending. A separate KPMG survey found that just 26% of companies have full visibility into their AI costs, while 22% have none at all and in some cases learn what they consumed only when the bill lands.
Gartner predicts that AI coding costs will surpass the average developer’s salary by 2028, driven by rising token consumption and vendors’ shift from seat-based to consumption-based licensing. According to the firm, nearly a quarter of technology leaders already spend $200 to $500 per developer each month on tokens for coding agents.
Gartner’s analysts argue that developers naturally favor speed and convenience over cost, so they recommend that organizations introduce token thresholds, escalation policies, and automated monitoring instead of relying on individual restraint.
The recommendations Gartner describe chime with how Wiv Air handles budgets in practice. Managers define policies that change access before employees run out of budget. When an employee reaches 70% of their allocation, for example, Wiv Air can restrict access to premium models while keeping lower-cost models available, so work continues and the remaining budget lasts longer. At the spending cap, the platform pauses the associated access through the provider’s API, with controls available at the employee or individual API-key level, and managers can reassign unused budget between employees and teams.
The broader AI-cost pressure is unlikely to ease. Gartner expects model pricing to keep climbing as providers recover infrastructure investment, and agentic workloads multiply the number of calls behind every task an employee hands off.
For CFOs, there needs to be a change in priority – moving away from what AI cost last quarter and toward who is allowed to spend what, on which model before tokens are consumed. Wiv Air promises to help businesses navigate this. Those who solve the AI cost dilemma will be able to widen AI access with confidence, while those that keep relying on monthly reports may find themselves rationing tools after the money has long dried up.
Wiv Air is available worldwide, including through AWS Marketplace.