There is a particular kind of morning that many people in organised work will recognise. You arrive with a sense of what you want to get done. Then you look at what you are actually able to do, and the answer is: not much, yet.
The supplier payment you prepared is waiting for sign-off. The discount you want to offer a long-standing customer needs a manager's approval. The purchase of a part to finish a job sits in a queue. The day you planned has quietly become a day of waiting, and the order in which things happen will be decided by when other people get to their approvals, not by you.
Nothing dramatic has happened. No one has been unreasonable. A system, and the rules built into it, is simply deciding what you can do today.
Autonomy is measurable, and unevenly spread
Autonomy at work is not only a feeling. It can be asked about directly. Eurostat's statistics on job autonomy, drawn from a 2019 module of the EU Labour Force Survey, asked workers how much influence they had over the content and order of their tasks. Across the EU, 49.6% reported some or a large influence. Almost one in three, 31.5%, reported little or none.
The split by employment status is striking. Among the self-employed, 83.0% reported some or a large influence over the content and order of their tasks. Among employees, the figure was 43.7%.
These are European figures, and they should not be read as a measurement of South African workplaces. They do show something general: being an employee often means that someone or something else sets the sequence of your day. Some of that is inherent in working inside an organisation. Some of it is a product of how the organisation's systems are designed.
The ILO has made a related point in its work on generative AI. It concluded that most jobs are more likely to be changed than replaced, and that the larger concern is the effect on job quality, including work intensity and autonomy. In other words, how technology is configured matters for how much control people keep over their work, not only for how many jobs exist.
How approvals shape a day
Approvals are one of the most direct places where system design meets personal autonomy. Take an illustrative case: a Durban logistics business with around forty staff, where every purchase above R2,000, every customer credit note and every change to a delivery schedule needs a manager's approval in the business system.
The rules were set years ago when the business was half the size and the owner wanted to see everything. Since then, prices have risen, volumes have grown and the same three managers now approve dozens of items a day between meetings and site visits. A dispatcher who could solve a customer's problem in five minutes instead spends the morning watching a pending status.
The mechanism has several parts:
- Thresholds that never moved. Limits set for a smaller business become tighter in real terms every year, so more decisions fall into the approval net without anyone choosing that.
- Approvers who cannot see the queue. If approvals arrive as scattered emails or notifications, the approver may not know that a person is blocked.
- No distinction between routine and exceptional. A repeat order from a known supplier goes through the same gate as an unusual one-off purchase.
- No information on who approves what. As described in when staff need to ask who approves what, people spend time working out the route before the approval can even start.
Mellorca has looked at the operational cost of this in why approval processes become bottlenecks and the capacity cost in approval queues that block capacity. The human side is less discussed.
The human tax
The tax here is autonomy, and it shows up in small ways. The person cannot choose the order of their own work. They cannot finish what they started. They may have the knowledge and judgement to make a decision but not the permission, and the system makes that gap visible every time they open it.
It would be wrong to claim that everyone experiences this as a loss. Some people prefer a clear rule that tells them when to check with someone else; it removes the weight of deciding. Others find it frustrating to be trusted with customers but not with a small credit note. Which experience applies depends on the person and the role, and it should not be assumed.
What can be said is that a poorly designed approval structure removes choices from people without anyone deciding that it should. The removal is a side effect of configuration, not a management decision that was weighed and agreed.
What could change
The answer is rarely to abolish approvals. It is to make them proportionate and visible.
- Review the thresholds. Ask which limits still reflect real risk and which simply reflect the business as it was. Adjusting a limit is a small configuration change with a large effect on who can act.
- Separate routine from exceptional. Pre-approved suppliers, standard discounts within a band, or repeat orders can be allowed to proceed, with exceptions routed for review.
- Move from approval before to review after where risk allows. For low-value, reversible actions, a weekly review of what was done can give the same oversight without blocking anyone.
- Make queues visible to approvers. A single list of pending items, with how long each has waited, lets approvers see the cost of delay.
- Publish the authority matrix. People should be able to see what they may decide themselves.
In the illustrative logistics business, raising the purchase threshold, letting dispatchers make schedule changes within agreed rules, and giving managers one daily approval list might clear most of the routine queue. The dispatcher's morning would then be about solving customers' problems rather than waiting for permission to do so.
Trade-offs
Approvals exist for good reasons, and any redesign should take those seriously.
- Control and fraud risk. Segregation of duties protects both the business and its staff. Widening authority without monitoring can create real exposure.
- Accountability moves. When a person may decide alone, they also own the consequences. For some, that is welcome; for others, it is a new source of pressure.
- Managers may lose visibility. If review after the fact is not actually done, oversight disappears rather than changes form.
- Rules can multiply. Trying to define every exception in advance can produce a matrix nobody understands.
A redesign that ignores these is not returning autonomy; it is shifting risk onto people without support.
Why it matters
The business reasons for better approval design are familiar: faster service, less idle time, fewer frustrated customers. The human reason is quieter. When routine decisions sit within a person's own authority, the shape of their day starts to belong to them again. They can decide what to do first. They can finish what they started. They can use the judgement they were hired for.
That is the dividend: not freedom from all oversight, but a fairer division between what the system decides and what the person decides. The test of a good approval design is not only how quickly items clear. It is whether the people working inside it can recognise their day as their own.
Sources and further reading
- <span class="mw-page-title-main">Job autonomy and pressure at work - statistics</span> (last verified 10 October 2026)
- Generative AI likely to augment rather than destroy jobs (last verified 10 October 2026)
Related Mellorca reading
- When the Founder's Identity Is Fused With Every Task
- Ten Open Tabs and Five Logins: What the Morning Sign-In Takes
- The Report Someone Rebuilds by Hand Every Friday