Voice AI: Beyond Transcription with Granola, CoLoop & EdgeTier
AssemblyAI16 June 2026Watch on YouTube
Part of series
Ep. 1 · Voice AI in de Praktijk
Experts en engineers bespreken hoe je robuuste voice AI-pipelines en agenten bouwt die verder gaan dan eenvoudige transcriptie.
View the seriesDescription
At Granola’s London office, we hosted a live panel on how voice AI is moving beyond transcription into structured outputs, insights, and action. Joined by Jonathan from Granola, Adrien from CoLoop, Shane from EdgeTier, and moderated by Ryan from AssemblyAI, we dug into the full voice AI pipeline: transcription quality, diarization, post-call vs. real-time tradeoffs, multilingual support, noisy audio, evaluation, and what it takes to turn raw conversations into useful product experiences. The conversation covered what actually matters when building with voice today—and what still needs to get better. ▬▬▬▬▬▬▬▬▬▬▬▬ CONNECT ▬▬▬▬▬▬▬▬▬▬▬▬ 🖥️ Website: https://www.assemblyai.com 🐦 Twitter: https://twitter.com/AssemblyAI 🦾 Discord: https://discord.gg/Cd8MyVJAXd ▶️ Subscribe: https://www.youtube.com/c/AssemblyAI?sub_confirmation=1 🔥 We're hiring! Check our open roles: https://www.assemblyai.com/careers ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ #MachineLearning #DeepLearning #VoiceAI #VoiceAgents #AIAgents #LLM #SpeechToText #AssemblyAI #ConversationalAI
What you'll learn
- Voice AI extends beyond transcription to transform conversation data into actionable insights and structured outputs.
- Diarization and transcription quality are critical components of a complete voice AI pipeline in production environments.
- Real-time processing and post-call analysis involve different technical tradeoffs depending on the use case.