Every enterprise on the planet has a learning and development strategy. The problem with that? Almost none of the plans are working. It's not because the intent is wrong: organizations genuinely want to evolve and grow their workforce.

Rather, the problem is the architecture paradigm they've bet on is fundamentally misaligned with how people actually learn and how AI is changing work. The de facto model is familiar: a monolithic learning management system (LMS) housing thousands of courses, compliance modules and certification paths. Employees log in, click through, pass a quiz and resume their work with a certificate that proves they spent 90 minutes consuming content that bore no relationship to their actual problems. This isn’t learning; it's performance management theater.

The problem no one wants to say out loud 

LMS platforms are designed around a school model — discrete courses, fixed curricula, periodic enrollment. That model makes sense when knowledge has a shelf life of years and jobs are stable enough to plan around. In the era of enterprise AI, those assumptions are no longer true. 

Related:Tips for successfully exiting AI vendor contracts

The half-life of a technical skill is now measured in months. Roles are being redefined faster than any curriculum committee can track. The fact that work itself is increasingly AI-augmented means the gap between "what the LMS trained you to do" and "what your actual job requires today" widens every quarter.

It isn't that employees won't learn: It's that organizations aren’t connecting learning to ways of working.

There’s a hidden cost to LMS-based instruction that rarely shows up in procurement analysis: It creates passive learners. When training is something you do separately from work — on a schedule, in a system, for a score — it trains people to receive knowledge rather than develop judgment. And judgment is precisely what AI-integrated work will require from us. It is also what AI cannot supply on its own.

What real-time, intelligent learning actually looks like 

The shift organizations need to make isn't incremental. It's operational and architectural. The learning layer must move out of the LMS and into the workflow, embedded in the tools, the decisions and the moments when capability gaps surface. This is where agentic AI changes the equation entirely. 

An AI agent doesn't wait to enroll an employee in a course about customer escalation handling. It observes that a rep is struggling with a specific objection pattern, surfaces the relevant coaching content in context, models a better response and tracks whether the outcome improved. 

Related:Before the next VMware: How CIOs prepare for vendor shocks

The feedback loop is real and immediate, not retrospective and nominal. The distinction that matters: Traditional learning and development (L&D) asks, "Did the employee complete the training?" Agentic L&D asks, "Did the employee's performance change?" One measures activity. The other measures the only thing that matters: a business outcome.

Agentic systems can also do something no static curriculum can: personalize at scale without human overhead. They can identify which individuals on a team carry which capability gaps, map those gaps to the organization's near-term strategic needs and route targeted development interventions. Those are not generic course recommendations, but specific, contextual nudges delivered at the moment of need. This isn't a feature enhancement to an LMS. It's an entirely different learning paradigm.

The workforce transformation imperative

The organizations that successfully embed AI into their operations over the next three to five years are not the ones that deploy the most AI tools. They're the ones that build a workforce capable of working with AI fluently — people who understand where to apply it, where to resist it and how to audit its outputs.

Related:Priceline CTO prioritizes engineers able to 'hold a room and a roadmap'

That capability cannot be delivered by a course catalog. Instead, what's required is a continual development operating model; one where learning is ambient, feedback is immediate, skill gaps are visible to managers and individuals in real time, and AI agents operate as ever-present, on-the-job coaches. 

Enterprises that make this transition will gain a critical structural advantage. Their workforce compounds. Every month of AI-embedded, real-time development builds on the last. Meanwhile, organizations still running annual compliance cycles and self-paced e-learning libraries are not just standing still; they're falling further behind a rapidly accelerating curve.

What leaders must do differently

None of this happens by licensing a new platform, AI or otherwise. The architectural shift requires three things that are harder than procurement:

The organizations getting this right aren't waiting for their LMS vendors to build AI features. They're asking "how do we build a system that makes our people better at their actual jobs, in the flow of real work?" That question has an answer. It just doesn't fit in a course catalog.

About the Author