The L&D industry’s AI problem isn’t adoption - it’s knowing how to use it for real impact
Access Learning report reveals top challenges for 2026, including the role of AI, skills and human expertise.
The landscape of AI in the workplace has transitioned from an initial focus on safety and compliance to a new stage beyond the “honeymoon phase” of adoption. Rather than simply layering AI onto existing structures, organisations are now embedding technology directly into workflows to improve outcomes. But, as organisations move from adoption to application, a harder question is emerging: is AI being used to drive real performance impact, or simply to produce more content, faster?
The Future of L&D 2026 report from Access Learning, part of The Access Group, brings together perspectives from leading industry voices to explore how the role of L&D is changing in response to these shifts.
While AI was initially introduced to speed things up rather than solve a clearly defined performance problem, the connection to real work, both in training and application, has become less impactful. Now, focus is no longer on whether to use AI, but on whether it’s actually making a difference.
Andy Lancaster, chief learning officer at Reimagine People Development, highlights how this shift is translating in practice and the key weaknesses in how organisations have built their AI skills.
“Previously, the conversation was how do we adopt AI safely, or what tools can we use, and how can we test those. Now it’s moving far more to how AI can measurably improve performance. It’s not about layering AI onto existing processes; it’s about redesigning those processes so AI sits at the heart of them.
“We’ve seen a slight kick back to AI literacy training: there was a push to train everybody up in basic AI skills, but we’re now recognising that it’s too generic, and doesn’t really influence performance. There’s much more thinking now around ensuring AI is contextual, as AI knowledge alone is not enough.”
Tom McDowell, founder and principal consultant at Evolve, added:
“Organisations that are struggling tend to be treating AI adoption as a technology project. Without a clear performance problem to anchor the technology, AI becomes a solution searching for a question – and that’s where you see the pattern of generating more content nobody needs, slightly faster than before.”
In addition to this, learners are becoming more sceptical about AI-generated information, with Sarah Mian, director of learning content at Access Learning, saying:
“There’s become this instinct now where people are asking, ‘Is that AI-generated?’, and the minute they think it is, it suddenly becomes less valuable.”
Getting AI to drive performance also depends on whether it reaches everyone in the workforce, with the study pointing out that the gap in learning accessibility remains a major problem. Experts warn that unless AI is designed with a deep understanding of these needs, it could actually make exclusion worse, especially since so much existing training material is still stuck in old, non-inclusive formats.
Susi Miller, founder of the eLa1000 eLearning Accessibility Assessment, highlights a common misconception within organisations, fuelling a lag in inclusive content production:
“Many organisations still significantly underestimate the number of people in their workforce with disabilities and access needs, often assuming it is around 4 to 7 percent, when the reality is closer to 25 percent. As AI platforms and tools evolve, there is a real risk that learning moves forward in a way that leaves people behind.”
However, this can be fixed. When designed effectively, AI can support more adaptive learning experiences by adjusting content, format, and delivery in response to individual needs.
Looking forward, progress will need to depend on a more focused and practical approach. This starts with a stronger emphasis on performance – working more closely with the business to identify where performance needs to improve, and ensuring learning is aligned to those outcomes.
The right digital learning software can already support these shifts, enabling organisations to deliver more personalised, performance-focused and accessible learning experiences at scale.
Elliot Gowans, General Manager at Access Learning, said:
“The fundamentals of good L&D haven’t changed. Understanding what people need, designing learning that’s worth their attention, being close enough to the business to know whether it’s making a difference: these things matter as much as they ever did. What’s changed is the urgency around all of it, and what’s coming next will change it further.
“We’re already seeing systems that can handle the routine groundwork of L&D: identifying who needs what, pulling together the right content, getting it ready to deploy. That frees up L&D professionals to focus on the work that needs human judgement. At Access Learning, that is where our main focus is: giving teams the tools to spend less time on administration and more time on the decisions that make a real difference.”