Skip to content

Platform · Policy

Define what agents may do

Authorization starts with rules your risk team can read. GREENLIGHT ships a native rule library for consequential agent actions and lets you add your own — every rule versioned, testable and promoted deliberately from test to production.

Part of GREENLIGHT

Why it matters

The problem Policy solves

Most agents run on permissions written for people or service accounts. Those permissions say what a system can reach, not which actions are acceptable, at what size, and when a person must decide. Policy is where you write that down, once, for every agent.

Questions it answers

  • What are our agents allowed to do, and up to what limit?
  • Which actions always need a person?
  • What can never be switched off, even by an administrator?

Where it sits

Policy in the GREENLIGHT loop

Every consequential action moves through the same loop. Evidence from each decision informs the next policy.

How it works

Policy, step by step

  1. 01

    Start from the native library

    Rules for payments, account changes and data movement are active from the first day.

  2. 02

    Add your organisation’s rules

    Write them in plain language and set the thresholds that match your appetite for risk.

  3. 03

    Preview the impact

    See which past actions a change would have affected before it goes live.

  4. 04

    Promote with approval

    Move rules from UAT to production through an approved, versioned change.

What’s included

The 3 modules of Policy

Policy · 01

Native policy library

A library of 550+ native rules for consequential actions — payments, refunds, account changes, data movement — active from day one.

  • Grounded in the obligations regulated firms answer to
  • Covers money movement, identity changes and data handling
  • Updated as guidance evolves

Policy · 02

Policy Studio

Write organisation rules in plain language, see what they would change before they go live, and promote them from UAT to production with approval.

  • Plain-language authoring
  • Impact preview before activation
  • Versioned, with promotion between environments

Policy · 03

Protected floor

Native protections can be tightened, never weakened. Where a threshold is designed to be tuned, you tune it — the floor itself stays in place.

  • Organisation rules cannot override protected rules
  • Adjustable thresholds where the rule allows
  • Every change recorded

In practice

Setting a refund limit

Operations want agents to settle refunds up to S$500 on their own, with anything above going to a team lead. They set the threshold in Policy Studio, preview how many refunds last month would have been held, and promote the rule once risk approves it.

Illustrative scenario.

Use case

Payments & refunds

Who uses it

  • Risk & compliance set the rules
  • Operations tune the thresholds
  • Engineering promotes changes between environments
  • One rulebook for every agent

    The same rules apply whichever framework or vendor built the agent.

  • Changes you can stand behind

    Every rule is versioned, previewed and approved before it applies.

  • A floor that holds

    Protected rules can be tightened but never quietly weakened.

See it working

See Policy on your own workflow.

We connect one agent workflow in shadow mode and walk you through every decision GREENLIGHT would make, so you can judge it on your own traffic.