The wrong response to technological change is universal upskilling. Different people need different depths of knowledge—and deciding who needs what is a leadership responsibility.

This piece is a companion to The Learning Triage, which helps individuals decide what to learn next. This one addresses the organisational version of the same problem: deciding who needs depth, who needs literacy, and who can safely wait.

Every technology shift eventually produces the same meeting.

A new platform, framework or capability appears. Someone explains why it matters. The conversation turns to skills, and a familiar question follows:

How do we train everyone?

It sounds responsible. It is usually the wrong question.

Not everyone needs to operate the technology, and not everyone will make decisions about it — plenty of people’s work will remain largely untouched.

Giving all three groups the same training may look inclusive, but it wastes attention, creates shallow knowledge and hides the harder decision: who actually needs to know what?

The problem is not access to learning. We have more courses, demonstrations, summaries and AI-generated explanations than anyone could consume.

The scarce resource is attention.

The scarce resource is attention.

New does not mean equally important

Technology encourages a false equality between new things.

A minor feature update and a genuine platform shift arrive through the same channels, occupy a similar amount of screen space, and carry the same implied warning: understand this immediately, or risk falling behind.

Newness starts to look like importance. But developments do not affect everyone equally.

Some change the work people perform every day. Some alter the environment in which decisions are made. Many make no meaningful difference to a particular role—at least not yet.

Treating every development as equally deserving of attention is not curiosity. It is a failure to prioritise.

A better capability strategy separates learning into three levels.

1. Learn deeply

Deep learning belongs with the people who will build, operate, assure or remain accountable for the capability.

Depth is more than recognising terminology or repeating the accepted view. It comes from practice: using the technology, encountering its limitations and developing judgement when the instructions no longer fit the situation.

If AI changes how a team designs, builds, tests or supports a service, that team needs more than an awareness session. Its members need to understand where the technology helps, where it fails, what good practice looks like and what new risks it introduces.

They need somewhere real to use it.

This matters because deep learning is expensive. It requires concentration, repeated practice and a willingness to be temporarily incompetent. It also takes attention away from something else.

That cost is precisely why depth must be selective. Before assigning it, ask:

Will this person repeatedly use the capability, make technical judgements about it or own the consequences when it fails?

If the answer is yes, a saved link or one-off course is not enough. The learning needs protected time, access to real work and support from people who already possess the capability. Without a practice surface, training becomes content consumption.

That cost is precisely why depth must be selective.

2. Understand broadly

Not every important development requires hands-on mastery. Sometimes people need a reliable map without needing to build every road.

A technology may affect customers, suppliers, colleagues or strategic choices without becoming part of someone’s daily work. Broad understanding is appropriate when a person needs to know:

• what the technology does;
• what it makes possible;
• where its boundaries and risks are;
• which claims should be challenged; and
• when deeper expertise needs to enter the conversation.

Done well, this is disciplined situational awareness, not the superficial pass it can look like from the outside.

Senior leaders do not need to become machine-learning engineers to govern an AI-enabled organisation. They do need enough understanding to challenge an unsupported proposal, recognise when a demonstration is being mistaken for an operating capability and ask who remains accountable when a model is wrong.

A product owner may not need to configure the platform but must understand how it changes cost, risk and customer experience. A procurement team may never operate the system but still needs enough knowledge to test a supplier’s claims. The distinction that actually matters underneath all of this:

Does this person need to operate the capability, or make sound decisions around it?

Confusing those two needs creates familiar problems. Leaders pretend to possess technical depth they do not have, while experienced professionals spend months mastering tools they will never use.

Broad understanding earns its keep when the purpose is clear. Presented as universal expertise instead, it curdles into theatre.

3. Ignore confidently

This is the category organisations find hardest.

Ignoring a subject can sound complacent. In technology, it is easily mistaken for resistance to change.

But confident ignorance is not incuriosity. It is a conscious, revisable decision that a subject does not currently affect someone’s work, decisions or responsibilities.

The word currently matters. Good triage always includes a return condition.

A team might ignore a technology until a customer asks for it, a regulatory expectation changes, a credible use case emerges or the capability moves from demonstration to durable practice. Until one of those signals appears, attention stays elsewhere.

Here’s a cleaner test than “do they need to know this”:

What would actually happen if this person knew nothing more about the subject for the next six months?

Often, the honest answer is: nothing. That is useful information. The purpose is not to celebrate ignorance. It is to stop manufacturing learning activity where no operational need exists.

The purpose is not to celebrate ignorance. It is to stop manufacturing learning activity where no operational need exists.

The universal-training trap

When a new capability becomes strategically fashionable, organisations often respond with scale before specificity.

Everyone receives an introductory course. Completion rates are reported. Capability dashboards turn green. A leadership presentation announces that thousands of people have been “upskilled”.

Yet little changes in the work.

Teams attend courses before they have a system on which to practise. Leaders learn vocabulary without gaining the confidence to challenge decisions. People whose roles are unaffected sit through generic material and forget it within weeks.

I’ve sat in the room where the completion-rate slide gets presented as a win. Nobody in it could tell me what had actually changed in the work.

The organisation has measured exposure, not capability.

Universal training is attractive because it avoids making distinctions. Nobody has to decide which roles genuinely need depth, which need literacy and which do not need attention yet. But those distinctions are the work.

Before launching a broad learning programme, leaders should be able to answer:

• Who will build, operate or assure this capability?
• Who will make decisions around it?
• What decisions will better knowledge improve?
• Where will deep learners practise?
• Who can safely wait?
• What signal will tell us that their need has changed?

If those questions do not have clear answers, the training programme is probably premature.

Attention is a portfolio

A healthy capability portfolio avoids two extremes: chasing everything, which produces constant movement without accumulated expertise, and dismissing everything, which protects established knowledge long after the work has changed.

There’s no perfect balance to strike between the two — the right shape is a deliberate asymmetry.

At any one time, an organisation should have:

• a relatively small group developing genuine depth;
• a wider group maintaining enough understanding to make sound decisions; and
• a much larger group that has consciously been given permission to focus elsewhere.

The final group will usually be the largest. Read that the right way and it isn’t a sign the organisation is falling behind — it’s a sign it has priorities.

Capability is not evenly distributed

Organisations sometimes describe capability as if it were a substance that should be spread evenly across the workforce. Real capability does not work like that.

It forms around responsibilities, relationships and places where knowledge can be applied. Some people need depth because the work depends on them. Others need enough literacy to ask better questions. Many need access to expertise rather than possession of it.

The objective isn’t for everyone to know the same things — it’s for the organisation to have the right knowledge in the right places, with reliable routes to reach it when needed.

That requires more than a course catalogue. It requires role design, communities of practice, access to specialists, time for experimentation and clear accountability when technology influences a decision.

The leadership decision hiding inside learning

The internet will continue to produce more useful material than anyone can absorb. AI will make that material easier to create, summarise and teach.

Access to knowledge is becoming cheaper.

Attention is not.

The valuable skill is no longer simply learning quickly. For individuals, it is deciding what deserves finite capacity. For leaders, it is deciding where depth must exist across the organisation—and protecting people from learning demands that add no value.

Learn deeply where people operate the capability or own the outcome. Understand broadly where people make decisions around it. Ignore confidently where it changes neither—and define the signal that will bring it back into view.

Not everyone needs to learn everything.

Not everyone needs to learn everything. A capable organisation is not one in which everybody knows the same things. It is one that knows what deserves depth, where that depth belongs and how everyone else can reach it when the work demands it.