A HR director told me last week that their employees are “trained out.” Between mandatory compliance modules, leadership development, technical certifications, and now wave after wave of AI training, people are hitting a wall.

The irony isn’t lost on anyone: we’re trying to help employees adapt to change by overwhelming them with more change.

How We Got Here

The pressure is understandable. Every second article about AI includes something about “the skills gap” or “workforce transformation.” Boards ask CEOs about AI readiness. CEOs ask CHROs. CHROs ask L&D.

And L&D, wanting to demonstrate responsiveness, launches training programs. Often quickly, sometimes without clear objectives, frequently with mandatory participation.

The result: employees who’ve been through three ChatGPT workshops, two “AI fundamentals” courses, and a mandatory “AI ethics” module—but still aren’t sure how AI applies to their actual job.

What Fatigue Looks Like

It’s not always obvious. People attend the sessions. They complete the modules. The LMS shows green checkmarks.

But actual adoption remains flat. Survey after survey shows employees feel less confident about AI, not more. The training created compliance without competence.

The warning signs:

• Declining engagement metrics in AI-related learning content
• Feedback comments about “another AI thing”
• High completion rates paired with low application rates
• Managers asking for exemptions or deferrals
• Training becoming a checkbox rather than a skill-building exercise

The L&D Response Making Things Worse

When initial training doesn’t produce results, the common L&D response is more training. Maybe the content wasn’t good enough. Maybe we need a different vendor. Maybe we need more advanced modules.

This doubling down amplifies fatigue. Employees see another calendar invite for another AI session and their eyes glaze over.

What’s actually needed is less training, not more—but better targeted.

A Different Approach

The organisations I’ve seen navigate this well do a few things differently.

Role-specific pathways. Instead of blanket “AI literacy” for everyone, identify which roles actually need which capabilities. A marketing coordinator needs different AI skills than a financial analyst. Training everyone on everything wastes time and attention.

Just-in-time learning. Deploy training when people are about to use a tool, not months before. A 15-minute refresher before someone starts using an AI writing assistant is more effective than a comprehensive course they completed six months ago.

Integration over isolation. Stop running AI training as separate programs. Embed AI capabilities into existing workflow training. If you’re teaching someone a new CRM, include how AI features work within that CRM. Don’t make them take a separate course.

Permission to skip. Let people self-assess and opt out of training they don’t need. Some employees are already proficient. Forcing them through beginner content breeds resentment and wastes everyone’s time.

The Manager Bottleneck

Here’s something L&D often underestimates: managers are the bottleneck for application.

Employees can complete every module, but if their manager doesn’t support applying new AI skills in actual work, nothing changes. And many managers are themselves overwhelmed, unclear on AI strategy, or skeptical of the technology.

Training needs manager enablement as much as employee content. Help managers understand what skills their teams are learning and create opportunities for application.

Measuring Actual Impact

Stop measuring training completion and start measuring capability application.

After an AI training program:

• Are employees actually using AI tools more?
• Are they using them effectively?
• Has productivity or quality improved in areas the training addressed?

These are harder to measure than completion rates. They also actually matter.

AHRI has some useful frameworks for learning impact measurement. Worth reviewing if your current metrics are purely activity-based.

What Employees Actually Want

When you ask employees what AI support they need, it’s rarely more formal training. They want:

• Time to experiment with tools
• Peer support and knowledge sharing
• Clear guidance on what’s allowed and what isn’t
• Examples relevant to their actual work
• Permission to make mistakes without career consequences

These needs don’t require training programs. They require culture and management practices.

The Path Forward

L&D teams face a difficult position. There’s genuine pressure to develop AI capabilities, but the methods many are using are backfiring.

The answer isn’t abandoning AI development. It’s being more strategic about what training is actually necessary, for whom, and when. It’s trusting employees to direct some of their own learning. It’s focusing on application environments, not just content delivery.

Less training, better targeted, with more support for actual application. That’s the path through AI training fatigue.