There is a meeting happening right now in a boardroom somewhere. A CEO just returned from a conference. They watched a demo. They are excited. And soon after, the learning team receives a message that starts with: "We need to be using AI in our training."
Sound familiar? Learning & Development teams across industries are navigating real pressure to adopt AI — fast. Executives see the headlines. They see competitors launching AI-powered onboarding. The expectation lands on L&D to deliver the same results, often without a clear brief on what problem is actually being solved.
The problem is not the ambition. The problem is that most organisations are rushing to the tools before they understand the problem they are actually trying to solve.
Three AI Adoption Patterns Worth Reconsidering
After working with enterprise learning teams across industries, I keep seeing the same three patterns play out. Each one feels like progress in the moment — but none of them reliably move the needle on learning outcomes.
Trap #1
The Content Factory Trap
Using generative AI to produce more training content faster. The result: a bloated library nobody uses, in a format nobody asked for. Speed without strategy is just expensive noise.
Trap #2
The Shiny Tool Trap
Purchasing an AI authoring platform because the demo was impressive. Six months later, the platform is underused and the team does not know what good learning design looks like — AI-assisted or otherwise.
Trap #3
The ROI Avoidance Trap
Celebrating how fast AI helps you build modules, while quietly ignoring whether those modules actually change behaviour or performance. Completion rates are not outcomes. They are activity metrics.
The Uncomfortable Truth About What AI Cannot Do
Here is what nobody at the AI vendor conference will tell you: AI can generate content at scale. What it cannot do — at least not meaningfully — is deeply understand your learners' daily workflow constraints, their existing knowledge gaps, their psychological readiness to change, or the specific business performance outcomes that matter to your organisation.
"AI accelerates production. It does not replace strategy. Building content faster only matters when you are building the right content in the first place."
The highest-performing L&D teams I have worked with do not use AI to produce more training. They use it to produce better-targeted training, faster. There is a critical difference — and it starts long before you open any authoring tool.
Three Principles for AI-Driven L&D That Actually Works
1. Start with the performance gap, not the AI tool
Before selecting any AI platform, ask the questions that matter: What specific behaviour are we trying to change? What does "good performance" look like in this role, six months from now? Where does the current gap actually sit — is it knowledge, skill, motivation, or environment?
If you cannot answer those questions clearly, no AI-generated module will help. The tool is not the strategy. It never was. But AI makes this mistake easier to make and more expensive to recover from.
2. Design for the workflow, not the LMS
The future of workplace learning is not a course library. It is learning embedded where work happens — in the ERP system, the CRM, the service desk tool, the onboarding portal. AI makes this far more achievable through intelligent content personalisation, contextual performance nudges, and on-demand support delivered at the moment of need.
Organisations that focus solely on completion rates inside an LMS are measuring activity, not impact. Designing for moment-of-need performance support is what builds genuine business outcomes — and that is where L&D can demonstrate its real strategic value.
3. Protect the instructional design layer
I will be direct here, because I see this mistake accelerating: AI can draft a script. It cannot build a learning arc. It can summarise content. It cannot sequence a skill-building journey that accounts for prior knowledge, cognitive load, and transfer to the job.
The instructional design layer — the human expertise that turns information into capability — is more strategically valuable right now than it has ever been. Not despite AI. Because of it.
What High-Performing L&D Teams Are Actually Doing
- Auditing their content inventory before buying new AI tools
- Using AI for first-draft scripting, then applying ID expertise to shape the learning arc
- Partnering with instructional designers to set quality standards for AI-generated output
- Measuring performance outcomes — not module completions
- Running small, well-designed experiments before scaling any AI solution
What Should You Actually Do This Quarter?
If your organisation is navigating AI adoption in learning and your stakeholders are already asking for a strategy, here is a practical starting point:
- Audit before you buy. Map your current content. What is actually being used versus what is collecting digital dust? AI will multiply your output — make sure the strategy behind it is sound before you scale it.
- Run one rigorous experiment. Pick a high-impact learning need. Design a solution using AI for content drafting and human expertise for instructional architecture. Set a clear performance metric before you launch. Use the result as your proof of concept — and your boardroom story.
- Reframe the conversation upward. When your CEO asks about AI, do not just show them a demo of faster content production. Show them a roadmap that connects learning investment to measurable performance outcomes. That is the conversation that elevates L&D from cost centre to strategic function.
- Invest in design quality, not just tool quantity. The organisations winning with AI in L&D are the ones who pair strong instructional design capability with AI productivity tools — not the ones who replaced their designers with a subscription.
Final Thought
AI is not coming for L&D. It is coming for the version of L&D that was already struggling — the content-heavy, completion-rate-focused, LMS-dependent model that was not delivering business impact anyway.
For learning teams willing to lead with strategy, invest in design quality, and use AI as a force multiplier rather than a shortcut — the next five years represent an extraordinary opportunity to redefine what L&D impact looks like.
The question is not whether you will use AI. The question is whether you will use it well enough to matter.
