Delegation of authority: hand off decisions without losing control
Delegation fails when a decision has no limit. Give every recurring decision one owner and a number.
Most owners I work with have tried delegating. They hired a good manager, said "go ahead, it's yours," and a month later the approvals were back on their desk.
The manager wasn't the problem. Nobody told them which calls they could make or up to what amount. So they did the safe thing: they asked.
For a long time that doesn't feel like a bottleneck. It feels like leadership. Someone hits a wall, you unblock them, everybody moves on with their life. But if that happens 15 times a day, you are the single point of failure. And the team learns to wait for you. I call it learned helplessness, and it's nobody's fault. It's what happens when the limits live in one person's head.
Telling a manager to "use your judgment" doesn't hand anything off. The limits you apply in your head are the one thing they don't have. So they guess and get corrected, or ask every time. Then the owner starts checking things personally, which is the exact behavior that taught everybody to wait in the first place.
Start with refunds, discounts and overtime
You don't need to map every decision in the company. Start with refunds, discounts and overtime. They come up every week and touch money. Each one looks too small to write down. Together, they're a big part of why the owner can't step away for two weeks.
Put a number on each role
Each recurring decision gets one owner and a limit. Below the limit, that person decides and moves on. Above it, the decision goes up one level.
For example:
- A sales lead approves discounts up to 10%.
- A service manager approves refunds up to $500.
- A shift supervisor approves up to 5 hours of overtime per person per week.
- On anything over $5K, the owner has the final say.
Your numbers will be different. What matters is having one.
The same works for collections. In a demo I use, an agent sends overdue reminders and logs the replies, then stops. The next step touches money, so it goes to a real person, let's call him Tom, and his job is to approve anything past 45 days. That gate is there because the call has consequences, and somebody has to own them.
The owner sees exceptions weekly, not every approval
Once the limits exist, the owner stops approving and starts reviewing. Once a week, one list: what went over a limit, who decided, what happened.
If the same exception shows up three weeks in a row, the limit is wrong. Change the number and keep the decision where it is.
This is how the owner keeps control. Fewer decisions reach them. The ones that do are the ones that matter.
Write it down once, review it quarterly
One page. The decision, the owner, the limit, who takes the exception. If it doesn't fit on a page, start with fewer decisions.
Then look at it every quarter. Prices change, people get promoted, volume grows. A limit set for last year will slow the team down or let too much through.
Before you write anything, take five recurring decisions your team makes every day. For each, ask yourself: could someone who isn't me make this call correctly without calling me? Do they understand the thresholds I apply before making a decision? Do they know when to escalate versus just handle it? Every no is a line on that page.
This comes before AI
An AI agent has the same problem a new manager has, without the habit of asking. If it doesn't know the limit, it will guess. And it will sound very confident doing that.
Every agent runs until it hits a gate a person still owns. The gate only works with a number on it and a name behind it.
In SCOPE™ by Efficiency Sprint, this is the Ownership state: people run it reliably. It comes before any agent touches the work, because you can't hand an agent a decision you haven't handed to a person yet.
Efficiency Sprint makes companies operationally ready for AI. In one engagement, $1.2M of $1.4M in overdue receivables was collected in three months, once one person owned receivables.
You always need to know who is responsible. And it can't be AI.
Luba Inz is the founder of Efficiency Sprint, which makes companies operationally ready for AI through consulting and the EfficiOS subscription.