When the LMS Dissolves: The Month L&D’s Stack Started Reorganising Around Agents

Hello, fellow Learning Futurists 👋

April–May 2026 is the month corporate L&D’s plumbing started to visibly rearrange itself around AI agents — and the month the first big institutional buyer publicly threatened to walk away.

Last month I wrote about the sector entering its accountability phase. This month, accountability got teeth. A major LMS shipped a product release that effectively unbundles learning content from the LMS. A vendor-funded but methodologically decent report quantified the AI-readiness gap inside enterprises — and it’s worse than most leaders think. And at California State University, faculty and students moved from grumbling about the $17 million OpenAI deal to formally petitioning the chancellor to walk away from it. Three different stories, but all asking the same question: who is this stuff actually working for?

This month, we’ll dive into:

The L&D platform layer reorganises around agents: what Docebo’s AgentHub launch, the MCP server play, and Bersin’s “platform of agents” thesis mean for L&D leaders still procuring traditional LMSs

The AI readiness gap, quantified: what the Docebo report and Microsoft’s Study and Learn GA tell us about why so much AI training is failing to land — and what scaffolded learning needs to look like in practice

The procurement reckoning at Cal State: why the faculty revolt against a $17m OpenAI deal is the most important enterprise-buying story of the month, regardless of which sector you’re in

Let’s go! 🚀