Someone still has to be responsible

An organisation receives three plausible recommendations about a difficult decision.

Each is clearly written. Each identifies risks. Each makes a persuasive case for a different course of action.

The recommendations have improved the discussion. They have not established who is authorised to decide, what happens if the decision fails or who will deal with the people affected.

Advice and responsibility occupy different places in the process.

A decision has a location

A recommendation can describe a situation. A decision changes it.

Someone commits money, changes a service, moves a deadline or asks other people to act. Those people experience the consequences whether or not the reasoning was eloquent.

As generating advice becomes easier, an organisation may acquire more analysis without becoming more capable of taking decisions.

The missing element could be information. It could also be authority, an unresolved disagreement or reluctance to bear the cost of choosing.

More recommendations cannot settle a disagreement about who gets to decide.

Responsibility can become diffuse

Imagine a hypothetical service team using an automated system to prioritise customer requests. A request is repeatedly placed at the bottom of the queue, and the customer suffers a meaningful loss.

The frontline team followed the process. The manager approved the system. The supplier designed it. The original decision to adopt it came from somewhere else.

Each explanation may be accurate. Together, they can leave the customer unable to find anyone who will resolve the problem.

This is a failure of organisation. The use of automation makes the chain more complicated, but it does not remove the need for an answer.

A service has to connect a complaint to someone with both the ability and the obligation to respond.

Authority must match the obligation

Naming a responsible person is insufficient if they cannot influence the system they are expected to oversee.

A meaningful role needs access to relevant information, time to examine it and authority to intervene. It also needs a realistic scope. One person cannot closely review an unlimited volume of decisions merely because their name appears on an approval screen.

Otherwise, oversight becomes ceremonial. The organisation has preserved the appearance of responsibility while removing the conditions that make it possible.

This is an important design question for any system that promises efficiency. What does the person supervising it need to notice? Can they notice it? What can they do when something is wrong?

Some decisions can be delegated

Responsibility does not require a person to make every routine choice manually.

A bounded process can make decisions within agreed limits. The organisation can examine outcomes, identify exceptions and provide a way to challenge mistakes.

The limits matter. A decision that is easy to reverse has a different character from one that produces lasting harm. A system performing a familiar task under stable conditions deserves different supervision from one operating in an unfamiliar situation.

The useful aim is to make delegation explicit. Which decisions are being delegated, under what conditions, and who remains responsible for changing those conditions?

The person who stays

There is a kind of value that only becomes visible after something goes wrong.

Someone understands the history, stays available, explains what happened and works through the consequences. They do not disappear when the original recommendation proves incomplete.

That continuity is difficult to express in a demonstration. It may look less impressive than speed or eloquence. Yet it is often what makes a service dependable.

As advice becomes more abundant, we should pay closer attention to the arrangements that turn it into accountable action.

A good recommendation helps someone decide. A responsible organisation makes sure the consequences have somewhere to go.