August 18, 2025
When Will AI in Learning Live Up to the Hype?
Artificial intelligence might be dominating headlines in the learning-tech space, but according to new research from the Fosway Group, most of the bold promises are still more potential than practice.
In fact, Fosway’s analysis shows that of all the shiny AI capabilities vendors are touting for the future, just 11% are live with customers today.
Right now, AI in learning is doing more heavy lifting in the efficiency department than in revolutionising how people learn. Companies are gravitating toward safe bets, solutions that fit neatly into current workflows and carry little risk. Think transcription services, automatic translation, or off-the-shelf content that can slot easily into existing platforms.
Fosway found the most widely adopted AI features in learning today include translations (text, voiceover, audio – 71%), AI-generated images or avatars (55%), automated translations (55%), and transcripts for video and audio (52%).
So far, so practical. But the research also points to where things are heading next.
The next wave of AI in learning won’t just trim costs, it’ll reimagine the experience altogether. Expect to see intelligent curation tools, generative AI used to design full courses, automated assessment creation, and conversational AI “study buddies” guiding learners through their training.
That’s where the real disruption lies: more personalisation, interactive practice with soft skills, and digital coaches that feel more like companions than clunky add-ons. Fosway calls this “edge advantage AI,” and while it’s still emerging, it’s where companies could unlock genuine innovation.
For learning leaders navigating the hype cycle, Fosway advises a practical mindset:
- Set a baseline – demand mainstream features as standard, don’t be dazzled by “exclusive” claims.
- Look ahead – ask vendors about their roadmap and how they’ll prepare customers for new features.
- Check references – especially for less common tools, get proof they work in practice.
- Experiment safely – pilot new features with vendors and learn by doing.
As lead analyst Myles Runham puts it: “AI is making a lot of noise in learning, but only 11% of promised capabilities are real for customers right now. The story so far isn’t about radical reinvention, but about streamlining and scaling what we already do. That in itself is a big shift—but the real transformation is still on the horizon.”
