Is L&D Building Ability While Ignoring the Capacity To Perform?

AI is now embedded in knowledge work. The question for Learning and Development (L&D) is no longer whether it will affect the function, but whether the function is prepared to change because of it.

Much of the response has remained familiar: use AI to create content faster, turn material into a podcast, produce a slide deck in minutes or personalise a course. These can be useful efficiencies, but they are not transformation. In many cases, they are the same L&D operating model with faster machinery attached.

In a Season 4 Episode 6 of AI for the Average Joe, Learning Consultant Andrew Jacobs offered a better way to frame the challenge: capability is capacity plus ability. L&D has traditionally concentrated on ability. AI is exposing how little attention we have paid to capacity.

AI has changed the work, not simply the content

If an employee can use AI to produce a report in an hour rather than a day, everything around that task is affected. Decisions may be needed sooner. Managers may have more material to review. Existing handovers and approval processes may become bottlenecks.

Yet many organisations have introduced AI while leaving meeting rhythms, workloads, decision rights and performance expectations untouched. People are producing more, but the wider system cannot absorb it. What gets described as productivity can quickly become more output, more pressure and more cognitive load.

If L&D responds by producing even more learning content, it risks adding to the problem it should be helping to solve.

Training does not create capability on its own

Someone can complete AI training, understand prompting and know how to scrutinise an output, yet still be unable to use those skills meaningfully. They may not know what is permitted, have access to the right tools, understand where human judgement is required or feel supported by their manager.

That missing piece is capacity: the space, permission, workflow and organisational conditions that allow someone to perform differently.

AI adoption is not gender-neutral

The conversation also explored the persistent gap in men’s and women’s deliberate use of AI at work. The issue is not necessarily that women are unaware of AI. Many are already encountering it through search, everyday tools and workplace systems. The gap is often in intentional, visible and confidence-led use.

Some women are still asking whether using AI is cheating, whether they are allowed to use it or whether their work will be judged differently if they do. Others may be using AI to manage workload, reduce cognitive load or create more space in an already demanding working life, rather than to compete visibly for recognition.

That matters because a generic AI training programme will not address those barriers. Leaders need to create permission, show practical value and make peer-to-peer learning visible. Women should not be left to solve this alone, either. If AI is reinforcing existing inequalities in confidence, visibility or progression, that is an organisational leadership issue, not a women’s problem.

Start with performance, not a course request

Before commissioning another programme, senior HR and L&D teams should ask:

What part of the work are we trying to improve?
How does the workflow change when AI participates?
Where must human judgement remain?
Which meetings, handovers or approvals are no longer needed?
What could prevent people from applying what they learn?

Those questions take L&D beyond skills delivery and into organisational performance. That is where the function now needs to operate.

L&D needs to become better at “heavy thinking”

Andrew described the combination of critical and creative thinking as “heavy thinking”. Critical thinking alone can leave organisations focused on risk and restriction. Creative thinking alone can generate endless pilots without enough scrutiny. Together, they allow teams to question assumptions, imagine better ways of working and test them responsibly.

For L&D, this means becoming less dependent on content as the default answer and more confident as an organisational consultant. The function can connect people who understand the work, uncover knowledge already held across the business, challenge outdated processes and help teams experiment safely. The organisation remains the knowledge engine; AI supports it rather than replacing it.

These are not peripheral future skills. They are central to whether AI adoption improves performance or simply accelerates existing dysfunction.

Agentic AI raises the stakes again

The move from standalone chat tools towards AI connected to workplace systems makes this more urgent. Once a tool can access organisational data, coordinate tasks or take action, a short session on prompting is nowhere near enough.

HR, L&D, IT, security, and business leaders need shared decisions on access, data, authority, oversight, and accountability. Nuance cannot be automated, and neither can responsibility. L&D has a legitimate role because every technical implementation is also a change in behaviour, judgement and work.

It is time to change what L&D measures

Completion rates and learner satisfaction will not tell you whether people can perform in an AI-enabled organisation. Better evidence includes improved decision quality, reduced rework, faster time to competence, fewer unnecessary handovers and greater confidence in knowing when not to use AI.

AI may take over parts of L&D production. That should not be treated as a threat. It creates room for the function to do the strategic work it has often claimed it wants to do.

The uncomfortable question is whether L&D will use that room wisely.

Ready to move beyond AI training as a one-off intervention?

We help HR and L&D teams build the skills, behaviours and organisational conditions needed for confident, responsible AI adoption. If your organisation is investing in AI but has not yet redesigned how people learn and perform around it, talk to us about the next stage of your strategy.