Tisha Chawla and Susheem Koul argue that model gateways cannot fully control agent costsAI agent cost governance attributes, limits and optimizes resource spending across complete AI agent runs. because runaway spending often emerges from loops, expanding context and sub-agent behavior inside a run. Their proposed control planeAn AI control plane observes AI activity and applies centralized policy without owning the application's primary task logic. attributes every model call to a run and user segmentAI cost observability explains which AI runs, users, models, tools and behaviors generated resource spending., records cost in a ledger and applies policies at the agent boundary.
Tisha Chawla and Susheem Koul describe lightweight boundary annotations that send inputs, outputs and cost signals to an out-of-band control planeAn out-of-band AI control plane observes and governs agent runs from a separate path outside the main task-execution flow.. A governor limits which interventions the platform may apply, allowing policies to compact context, reduce tool output or inject concise instructions before a hard budget cap terminates the runAn AI run budget cap is a hard spending or usage limit that stops an AI run from consuming more resources..
Tisha Chawla and Susheem Koul demonstrate preview, halt and steer modes on a two-agent research workflow. Their reported benchmark across open-source agent projects reduced average spend by about 78 percent while improving completion compared with simple throttling, and they propose using accumulated run data to refine or generate policies over time.
Watch on YouTube
