Why "use your judgment" is not a refund policy

"Use your judgment" is a common refund policy instruction that produces inconsistent outcomes because it gives agents no framework for borderline decisions. It sounds like trust - like you've hired good people and you're letting them do good work. What it actually is, in practice, is a decision made in advance to not make any decisions, and to pass the consequences of that down to your agents.

What "use your judgment" actually means to your agents

When a manager says "use your judgment" on refunds and exceptions, agents hear different things depending on who they are.

The generous agent hears: say yes when it feels right. They build warm customer relationships. They resolve things quickly. They also issue more credits than anyone intended, to anyone who pushes back with enough persistence. They're not doing anything wrong - they're doing exactly what they were told, through the lens of their own personality.

The cautious agent hears: don't get it wrong. They escalate more than they need to. They say no to borderline cases where a yes would have been the right call. They lose customers who deserved a better outcome. They're also doing exactly what they were told.

The result, across a team, is that customers get different answers depending on which agent they reach. Consistent service becomes a lottery. And nobody can tell from the outside which approach is correct, because there's no definition of correct - just judgment, applied differently by everyone.

Why managers say it

"Use your judgment" isn't laziness. It usually comes from one of two places.

The first is a genuine belief in agent autonomy. Good managers don't want to micromanage. They've hired people they trust, and they want those people to feel empowered to help customers without needing permission for everything. That instinct is right. The execution is wrong.

The second is conflict avoidance. Writing a policy means drawing a line. Drawing a line means some requests will be on the wrong side of it. "Use your judgment" defers that discomfort indefinitely - every borderline case becomes an individual call rather than a policy test. It feels more human. It's also more chaotic.

The three things a real policy has that "judgment" doesn't

A threshold. Something that tells an agent, before they even start the conversation, which decisions are theirs to make and which need a sign-off. A dollar value. A deviation category. A customer tier. It doesn't have to be complex. It has to exist.

A process for exceptions. When a situation falls outside the threshold, there should be a defined path - not a DM, not a tap on the shoulder, a structure. The agent requests. The manager decides. The reason is recorded. That's it.

A record. Every decision outside standard policy should be traceable. Who requested it, who approved it, what the reason was. Not because you don't trust people, but because patterns in that data will tell you whether your threshold is set right, which agents need support, and whether your policy is actually working.

"Use your judgment" has none of these. It's not a policy - it's the absence of one. We covered what that looks like at team scale in Why customer support teams need an approval process and The hidden cost of the shoulder-tap.

The difference between autonomy and ambiguity

Agent autonomy is real and valuable. Agents who feel trusted to resolve things make faster decisions, handle customers better, and stay in the job longer. The goal isn't to take that away.

The goal is to define the zone where autonomy lives.

An agent with a clear threshold - "you can approve refunds up to $X, anything above that needs a sign-off" - has more genuine autonomy than an agent with no threshold. They know exactly where their authority begins and ends. They don't have to guess. They don't have to ask permission for things that are clearly within their remit, and they have a clean path for things that aren't.

"Use your judgment" doesn't give agents autonomy. It gives them uncertainty dressed up as freedom - and it puts them in an impossible position every time a borderline case lands.

How to replace it

You don't need a fifty-page policy manual. You need two things written down.

First: what agents can do without asking. A dollar threshold. A list of situation types. Whatever makes sense for your team and your risk tolerance. This is the zone of genuine autonomy.

Second: what happens when something falls outside that zone. Where the request goes. What format it takes. Who sees it. How fast the decision comes back. How the decision is recorded.

For Intercom and Slack teams, Supportman Approvals handles the second part - the request comes from the Intercom inbox, the manager approves or rejects with a reason in Slack, and the decision is recorded automatically. The first part - the threshold - is yours to define. But once you've defined it, the structure exists to support it.

The result is a team where agents know what they can do, know how to get a decision on what they can't, and where every exception is traceable. That's what empowerment actually looks like.

Frequently asked questions