Execution pattern
Let support move fast without letting policy drift.
A refund, credit, or goodwill adjustment is issued to a customer.
Refunds are individually small and collectively enormous. Discretion is delegated widely and reviewed narrowly, so drift is only visible once it is a quarterly number.
The signals available at the moment of the action, and the rule that judges them.
Refuse refunds beyond the issuing agent's discretion, and escalate patterns that look like repeat abuse rather than one-off service.
The refund issues within discretion.
A larger or repeated refund waits on a supervisor.
The refund does not issue.
Fork one and change the thresholds. Available today on 3 platforms.
Lets the Zap issue refunds that sit inside policy and holds the rest — with the reason code attached to the record.
zapier
Wrap `send_refund`, `issue_payout`, or any tool that moves money — the inner tool only fires on an approve verdict.
langchain
Turns the refund approval that happens in a Slack thread into a signed verdict carrying the policy version and reviewer authority.
slack
Pauses a Zap the moment a customer's risk score crosses the line, instead of letting the automation keep firing.
zapier
Blocks an invoice or payment instruction from a domain that is not on the verified-vendor list before the Zap pays it.
zapier
This pattern is one answer to each of these.
The same five steps for every pattern in the library.
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.
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.
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.
Point the same policy at the system where the action actually originates — a checkout, an ERP posting, a Zap, an agent's tool call.
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.