This week: it's a quiet holiday week, so we're doing a mid-year stock-take — six months into the "AI changes everything" story, here's what's actually landed, what's stalled, and where the smart money is going.
1. What's actually happening
Buyers are unhappy with their LMS. Roughly 42% of companies say they're actively looking to upgrade or replace their learning platform — because the one they bought isn't delivering what they expected. The market's still growing fast (into the tens of billions), but the churn signal underneath it is the real story. (Svitla)
Agentic AI moved from hype to cautious pilots. In the latest survey data, ~27% of L&D teams say they're already exploring agentic AI, and another ~39% are interested but cautious. That's a lot of curiosity — and a lot of hesitation. (Synthesia, 2026)
Free AI upskilling is scaling. AWS's AI & ML Scholars program wrapped its Challenge phase on June 24, offering foundational AI education to up to 100,000 learners, with tracks like "Agentic AI Business Professional." Big tech is now a major provider of workforce AI literacy. (AWS)
The money is chasing "agents that learn on the job." NeoCognition came out of stealth with a 40Mseedforself−learningagents;ex−DeepMindresearcherDavidSilver′snewlabraised1.1B to build AI that learns without human data. Not L&D companies — but the tech that will reshape L&D tooling next. (TechCrunch · PRNewswire)
2. The half-time read
Six months in, the pattern is clear: AI has transformed how learning gets made, but not yet how it gets bought or governed. Content production is faster, cheaper, and everywhere. But the platform layer is lagging — buyers are frustrated enough to shop for replacements — and the agentic promise everyone demoed in the spring is landing as "interested but cautious" in the summer.
That gap between vendor ambition and buyer readiness is the whole game for the back half of 2026. The teams that win won't be the earliest adopters or the loudest skeptics — they'll be the ones running small, governed pilots now, so they actually know what works before the renewal cycle forces a decision.
3. By the numbers
~42% of companies actively looking to upgrade or replace their LMS.
~27% of L&D teams already exploring agentic AI; ~39% interested but cautious.
87% already using AI somewhere in L&D — but only ~9% scaling it across the org.
Up to 100,000 learners in AWS's latest free AI-literacy cohort.
The throughline: enormous adoption at the individual level, real hesitation at the system level.
4. Operator takeaway
If you have an LMS renewal in the next 12 months, start pressure-testing the skills-and-agent roadmap now — not the feature list, the roadmap. Ask each vendor: what can your agents actually do today, who reviews their decisions, and what's shipped versus promised? Half the market is about to go shopping at once; the teams that walk in with a tested pilot and sharp questions will get far better deals than the ones reacting under contract pressure.
5. On the radar
The billion-dollar bets on "agents that learn on the job" are worth watching closely. If AI agents can genuinely improve from real work rather than static training data, the line between "using a tool" and "being trained by it" starts to blur — and that's a fundamentally different model of workplace learning than anything in your LMS today. It's early, but it's the direction the capital is pointing.
Found this useful? Forward it to whoever owns training or enablement on your team — that's how this grows.
Got a take? Hit reply and tell us what's actually changing in your L&D world right now. We read every reply, and it shapes what we cover next.
— The LearningOps Brief team
