Loading…
Execution pattern
Prevent AI overspending
Cap what an agent can spend before the spend happens, not in next month's bill.
What executes
An agent calls a paid API, provisions a resource, or loops on a task that bills per token.
What normally goes wrong
Nobody notices until the invoice arrives. Agent cost is discovered retrospectively because the spend decision was never a decision — it was a side effect of a retry loop.
The decision
The signals available at the moment of the action, and the rule that judges them.
Signals
- Cumulative spend in the current window
- Cost of the specific call being attempted
- Budget delegated to this actor
- Whether the request is a retry of work already paid for
The gate
Refuse the call when it would take the actor past the budget delegated to it, and escalate when it crosses a threshold someone agreed to review.
ALLOW
The call proceeds and the spend is recorded against the actor's budget.
ESCALATE
The call waits for whoever owns the budget to approve the overage.
BLOCK
The call does not run; the dossier records the budget rule that stopped it.
Four stages left to right. Input signals — user or agent identity, context and data, tools and resources, spend and budget, policies, history and reputation, and environment — enter the Decionis Protocol. Inside, the protocol normalizes context and intent, evaluates policies and rules, runs a deterministic risk and policy engine, and explains the reason and evidence. It emits a verdict: ALLOW (authorized), ESCALATE (needs approval or more info), or BLOCK (high risk or out of policy). The verdict drives an enforcement action: allow, block, require approval, or throttle and limit. Every decision is recorded for audit, learning, and continuous improvement.
6 policies implement this pattern
Fork one and change the thresholds. Available today on 6 platforms.
Copilot Token-Budget Guard
Reserves projected model spend before the prompt is forwarded and restrains runs that would blow the project wallet.
copilot
AI Agency Pack
Token budgets, approved tool use, and brand-safe output — prove every client-facing agent action stayed inside the agreed rules.
langchain
AWS Agent Egress & Spend Gate
Opens sandbox egress only for the exact target the grant covers, and stops a run that would blow the account budget.
aws
SaaS Ops Pack
Cloud spend, discount, and entitlement gates — stop runaway burn, margin leakage, and unsupported entitlement changes.
sdk
Slack Discretionary-Spend Gate
Every spend approval ping above $5k carries the policy snapshot, the requester's authority, and a public verify URL.
slack
Zapier Spend-Approval Gate ($500+)
Require a named approver before any Zap releases spend over $500 — the approval becomes a signed verdict, not a thumbs-up.
zapier
Related concepts
This pattern is one answer to each of these.
What to do next
The same five steps for every pattern in the library.
- Simulate this recipe
Run the policy against a realistic action in the browser. Push it past what the rules allow and watch the verdict come back. No account.
- View the signed Decision Dossier
See exactly what was decided and why: the rule that fired, the evidence it read, the policy version in force, and an Ed25519 signature you can verify yourself.
- Deploy in shadow mode
Measure what the policy would have caught on your own traffic without touching the live path. Every recipe defaults to shadow, so the first deployment carries no execution risk.
- Connect a real workflow
Point the same policy at the system where the action actually originates — a checkout, an ERP posting, a Zap, an agent's tool call.
- Share the verified result
Publish the proof: a public verification link, an embeddable badge, a PR comment, or an anonymized shadow-mode finding. This is how the next person discovers Decionis.