The tension no one really talks about

Competency works best when the task, the tools, and the environment are relatively stable. You train someone to perform a task under defined conditions, and you assess whether they can do it consistently. That model holds up extremely well in a lot of industries.

Digital work doesn’t behave like that.

Interfaces shift. Platforms update. Processes get automated, then reworked, then replaced. AI tools change how tasks are approached altogether. The “right way” to do something today might not even exist in the same form six months from now.

That creates a mismatch.

We’re using a model designed to confirm performance in known conditions to prepare people for conditions that don’t stay known for very long.

That doesn’t make the model wrong. But it does mean we need to be careful about what we expect it to achieve.

Competency already does more than people think

It’s worth calling this out, because competency often gets unfairly simplified in these conversations.

In the VET context, competency isn’t just about doing a single task. It’s built around a broader idea of performance, including:

•        completing the task itself

•        managing multiple tasks

•        handling disruptions or problems

•        operating effectively within a workplace

There’s already an expectation that learners can adapt to some variation. Transferability isn’t absent from the model.

But—and this is the important part—that transfer usually assumes the situation is still recognisable.

You might be doing the same task in a slightly different context. The inputs might change. The setting might change. But the underlying structure of the work is still familiar enough that you can draw on what you already know.

That assumption doesn’t always hold in digital environments.

Where things start to break down

You see it most clearly at the task level.

A learner might be trained to use a system, follow a process, generate a report, or complete an online workflow. They can demonstrate that skill. They can pass the assessment. Everything looks solid.

Then the system changes.

The layout shifts. The buttons move. The workflow gets streamlined or expanded. A new tool replaces the old one. Suddenly the learner isn’t just applying the same skill in a new context—they’re trying to figure out what the task even looks like now.

What’s missing isn’t effort. It’s not motivation. It’s not even intelligence.

It’s the underlying understanding that allows them to reorient when the surface-level process disappears.

Managing tasks vs adapting to change

The same pattern shows up when you look at task management.

Learners can often manage a sequence of steps very effectively when those steps are stable. Log in, retrieve information, apply a template, submit a result. It’s structured. Predictable. Repeatable.

But digital workflows don’t stay fixed.

Automation removes steps entirely. AI introduces new decision points. Systems start doing parts of the job that used to be manual. The shape of the work changes.

At that point, it’s no longer just about managing tasks. It’s about making sense of how the task itself has shifted.

That’s a different skill.

When “something went wrong” isn’t obvious

One of the strengths of competency-based training is that it prepares learners to deal with problems. If something goes wrong, they’re expected to respond appropriately.

In digital environments, the problem is often harder to spot.

Outputs can look correct while being inaccurate. AI-generated content can be confident and coherent but still wrong. Automated systems can produce results that technically work but don’t align with what’s actually needed.

There’s no obvious error message. No clear signal that something has failed.

Instead, the learner has to decide whether to trust the output in the first place.

That’s not a standard contingency. It’s judgment.

Digital skills don’t sit neatly inside one role anymore

Another shift that’s easy to overlook is how digital skills cut across job roles.

In the past, it made sense to think about skills in the context of a specific occupation. But now, foundational digital behaviours—navigating systems, evaluating information, managing digital identity, working with AI—show up almost everywhere.

It doesn’t matter whether someone is in admin, retail, construction, aged care, or community services. The tools might differ, but the underlying demands are starting to look very similar.

That makes digital skills less like a discrete competency tied to a role, and more like a baseline capability that supports everything else.

So where does that leave competency?

Still essential.

This isn’t about replacing competency-based training. It’s about recognising where it needs support.

Competency does a very good job of answering the question: “Can this person perform the task under these conditions?”

What it doesn’t fully answer is: “What happens when those conditions change?”

That’s where capability comes in.

What capability adds

Capability is harder to define neatly, but you recognise it when you see it.

It’s the learner who can:

•        figure out a new system without needing step-by-step instructions

•        adapt when a process changes

•        question an output instead of just accepting it

•        transfer their understanding into unfamiliar situations

It’s not just about doing the task. It’s about understanding enough to keep going when the task evolves.

A framework designed for exactly this gap

This is precisely the problem that DigComp 3.0 was built to address.

Unlike traditional competency frameworks that describe what a person can do under defined conditions, DigComp 3.0 is a capability framework. Its levels don’t just measure task completion — they describe how someone operates: how independently, how adaptably, and how well they cope when conditions shift.

A Level 3 result tells you someone can work independently in familiar situations. A Level 4 tells you they can transfer skills to new tools and contexts. A Level 5 tells you they can problem-solve without a guide. These aren’t just performance descriptors — they’re capability indicators

That’s what makes DigComp a natural companion to VET competency frameworks rather than a replacement for them. Competency confirms what someone can do. DigComp tells you how resilient that performance is likely to be when things change.

Used together, they give a much more complete picture than either provides on its own.

Artificial Intelligence has made this impossible to ignore

If there’s one area where this shift is obvious, it’s AI.

You can absolutely train someone to use an AI tool in a very structured way. Enter a prompt, generate a response, complete a task. That’s assessable. That’s observable. That fits neatly into a competency framework.

But AI tools don’t stand still.

They change constantly. New tools appear. Existing ones behave differently over time. The same prompt doesn’t always produce the same result.

Consider a learner who has been trained to use an AI tool to summarise documents. They know the steps: upload, prompt, review, submit. But when the organisation moves to a different AI platform with a different interface and different behaviour, those steps no longer map. A learner with task-level competency is stuck. A learner with underlying capability — who understands what the tool is doing and why — can adapt.

That’s the difference.

Why this matters most at the start

At higher AQF levels, learners usually have enough experience to navigate this. They’ve built up a base of knowledge that helps them make sense of new tools and environments.

At lower AQF levels, that scaffold often isn’t there yet.

This is where training design matters most. Step-by-step instruction has genuine value — it gets learners through a specific assessment and builds initial confidence. The challenge is when that’s the only approach used. Without moments that build genuine understanding, learners can find themselves dependent on conditions staying the same.

And that’s the one thing digital environments don’t do.

A small shift with a big impact

The practical takeaway isn’t complicated.

Keep competency as the outcome. Start treating capability as the method.

That might mean:

•        explaining why something works, not just how

•        exposing learners to small variations instead of perfect repetition

•        building in moments where they have to think, not just follow steps

•        using tools like AI to explore understanding, not just generate answers

None of this requires rewriting training packages or abandoning compliance frameworks. It’s a shift in emphasis, not structure.

Competency-based training isn’t going anywhere. Nor should it. But in digital contexts, it’s no longer enough on its own.

If we want learners to succeed beyond the training environment, we need to build something alongside it. Something that helps them adapt, question, and keep moving when things change. That’s what capability gives us. And in 2026, that’s no longer a nice-to-have. It’s the difference between training that works on paper, and training that holds up in the real world.

ABOUT THE AUTHOR

David Cunning

David Cunning is the Programs Director of The Learning Resources Group. He has been in the VET sector for over 20 years and has spent more than decade managing the creation of training and assessment resources for over 300 units of competency. He was the driving force behind the LLN Robot System of assessing and supporting vocational education students across the country and has continued to develop solutions across multiple frameworks including DigComp for digital skills and capability. Dave has invested himself in understanding the industry by attaining his Certificate IV in Training and Assessment and also a Diploma of Vocational Education and Training and a Diploma in Training Design and Development.