AI has gone from fringe to mainstream in learning and development (L&D) in the blink of an eye. Most L&D teams are now experimenting with it in some form, whether that’s using AI to generate course content, automate admin, or augment the learning experience.
But despite this momentum, actual progress with AI beyond the experimental stage appears to be stalling.
According to Rise Up’s State of Learning Report, 62% of L&D leaders cite a lack of AI knowledge as their biggest barrier, while 55% say they don’t have the skilled people needed to implement AI-driven learning effectively.
So, while the ambition is there and the technology is ready, effective and sustained delivery remains just out of reach.
What’s striking is that, unlike a few years ago, these barriers aren’t about resistance or scepticism. L&D leaders broadly agree that AI has the potential to accelerate skills development and improve learning outcomes. Confidence is undermined by the maturity gap, where the promise of AI outpaces what people experience in day-to-day use.
Many teams are unsure where AI genuinely adds value, how to apply it responsibly, and how to measure whether it’s actually making a difference. AI-driven course generation is a good example here, as it is largely a one-shot solution to a single problem rather than a route to continuously developing company-wide capability.
To move forward, L&D needs to address these barriers head-on. This starts with understanding what’s really holding progress back and what can be done about it.
The knowledge barrier: Defining what the ‘AI knowledge gap’ actually is
When L&D leaders say they lack AI knowledge, they’re not talking about it in a data science or coding sense. It’s more fundamental than that. Most teams struggle to distinguish between different uses of AI in learning, such as automation, content generation, or course personalisation.
That means AI is often treated as a catch-all productivity tool rather than a capability-building technology. In the minds of these teams, it speeds up existing processes, but it doesn’t change how learning works or how quickly people develop the skills the business needs – and that’s where disillusionment sets in.
Without a clear understanding of where AI delivers meaningful learning value, teams default to safer, surface-level use cases. Content is produced faster, administration becomes lighter, but the underlying learning experience stays largely the same – so departments end up thinking “What’s all the fuss about?”
We need to start by reframing what AI knowledge really means for L&D. The priority isn’t technical mastery or prompt-based wizardry, but learning design fluency. Teams need to understand which learning problems AI is well-suited to tackle, where it can remove friction, and how it supports outcomes like faster progression to proficiency.
Building that understanding gives L&D the confidence to move beyond experimentation and start using AI to directly improve skill development, rather than simply making existing workflows faster.
The talent barrier: The perceived shortage of skilled people
“We can’t scale AI because we can’t find the right talent” – we’re hearing this more and more across the industry. A lot of L&D teams are small, already stretched, and unsure how to support new technologies without bringing in specialist expertise. And that uncertainty has become a blocker.
It creates the sense that meaningful progress depends on hiring AI experts or building entirely new capabilities from scratch, but that isn’t the case. Most L&D teams don’t need deep technical specialists to move forward, they simply need clearer ownership, better collaboration with IT or data teams, and tools that reduce complexity rather than add to it.
Upskilling existing L&D professionals to interpret data, define learning outcomes, and apply AI thoughtfully is far more sustainable than trying to compete for scarce technical talent.
The hidden barrier: Measuring the wrong things
Even when L&D teams overcome skills and resource concerns, progress with AI often stalls because success is measured in the wrong way.
Many organisations still rely on familiar indicators such as completion rates, satisfaction scores, or post-course feedback to judge whether learning is working. These measures say something about experience, but precious little about effectiveness. When AI initiatives are assessed through these lenses, it becomes difficult to prove their value or justify their use – especially at the board level.
Again, this ties back to confidence and how it breaks down. Without clear evidence that AI is actively improving things across the organisation, investment stalls and AI stays confined to small-budget pilot projects.
Teams can celebrate these one-off wins, but on an organisational level, the needle has barely moved.
A more useful approach is to focus on ‘time to skill’ – how quickly employees learn the skills they need. This offers a direct view of whether learning is accelerating performance, rather than just engaging workers in novel exercises.
An AI-driven course might be well received, but if the time-to-skill doesn’t shorten, it’s not having the desired impact.
AI is particularly well-suited to reducing friction in learning journeys, personalising learning pathways, and helping people reach proficiency faster. When L&D teams measure these outcomes, they gain a clearer picture of where AI is making a real difference and where it can be applied next.
Overcoming the AI adoption gap
L&D teams aren’t constrained by a lack of ambition or belief in AI. We know this because the technology has become ubiquitous, even if it isn’t being deployed to its fullest potential.
But they are being held back by a lack of AI confidence and understanding, unclear ownership, and outdated KPIs. The good news is that overcoming these barriers doesn’t require an army of AI engineers or data science experts – it simply requires more visibility, control, and collaboration across the business and a shift toward metrics that reveal the real picture.
Your next read: The AI-powered workforce: Where are you on the ‘three-loop journey’?
