ETS Global talk: when AI transforms speaking skills
1/10/2026 • Conférences
Invited by ETS Global, Miguel Membrado, cofounder and CTO of LinguoSphère, addressed a simple question that is rarely tackled head on: why do so many learners understand a language without daring to speak it at work?
The blind spot of language training
Most language programs focus on comprehension, grammar and writing. Yet the skill that unlocks a career, an integration path or a client relationship is spontaneous speech. It requires repetition, the right to make mistakes and immediate feedback, three things a traditional classroom struggles to deliver.
What conversational AI actually changes
- Practice volume: every learner can speak several times a week without mobilising a trainer.
- Context: conversations happen in real work situations, not abstract drills.
- Feedback: it covers clarity, job specific vocabulary and fluency, not only errors.
- Evidence: progress becomes measurable, giving training teams real reference points.
Assessment shifts too
Assessing speaking through a single test gives a snapshot. Observing dozens of authentic interactions gives a trajectory. That is where measurement becomes useful for employers: it documents progress rather than a frozen level.
Key takeaway
AI does not replace the trainer: it absorbs repetition, what humans do least well at scale, and gives trainers back the time for coaching. That principle is exactly what LinguoSphère is built on.