Guides
Field manuals for
agent governance.
Original, in-depth guides on AI agent cost control and authorization policies — how we think about spend caps, approval queues, policy simulation, forecasting, and lockdowns. No filler, no recycled takes.
How to Set AI Agent Spend Caps That Actually Stop Runaway Bills
9 MIN READPer-agent daily and monthly spend caps: the first defense against runaway AI bills. How to size them, hard blocks vs alerts, and failure modes teams miss.
The Human-in-the-Loop Approval Queue Playbook for Risky Agent Actions
8 MIN READNot every risky agent action can be denied outright. Routing rules, SLAs, mobile-friendly decisions and audit trails for human-in-the-loop approval queues.
Allowlists vs. Denylists: Writing Agent Action Policies That Hold Up in Production
10 MIN READAllowlists vs denylists for agent tool access: which default-deny posture holds up in production, and how to write action policies that survive real agents.
Forecasting AI Agent Spend: Predicting Your Monthly API Bill Before It Lands
9 MIN READPredict your monthly AI API bill before it lands. Burn-rate forecasting from live logs: sizing caps, spotting runaways early, and auto-throttle.
Step-Up Lockdown: Freezing a Misbehaving Agent Like a Bank Freezes a Card
8 MIN READFreeze a misbehaving agent the way a bank freezes a card: anomaly-triggered lockdowns where every action needs a human until the incident is released.
Put the theory to work
Every technique in these guides — caps, queues, simulation, forecasting, lockdowns — is a feature in the gateway, not just advice.