Shadow AI Spend
Shadow AI spend is AI cost that no budget, owner, or approval process accounts for, such as personal subscriptions, expensed seats, and API keys used outside a sanctioned path, and it usually surfaces only on an expense report or provider invoice.
Shadow AI spend is AI cost that no budget, owner, or approval process accounts for, such as personal subscriptions, expensed seats, and API keys used outside a sanctioned path, and it usually surfaces only on an expense report or provider invoice.
It is the cost side of shadow AI. Because nobody owns it, it is not attributed to a team or project, it is not forecast, and no budget caps it. Finding it starts with shadow AI discovery: an inventory of the AI tools people actually use.
Behest Radar shows which AI tools a machine uses and, with permission, reads their own local records: token counts for Claude Code, Codex CLI, and Claude Cowork; list-price cost for Claude models. That gives a team a first sense of the gap before deciding what to do. Bringing that usage under AI Token FinOps, the cost-control layer of the Behest control plane, turns it into attributed, budgeted spend.