News from Zoom, Gladia, Microsoft, PolyAI and others

Voice AI weekly digest

Top Updates 💪

• Gladia launches Solaria multi-lingual speech recognition model (VentureBeat)

• Voice AI platform Phonic gets backing from Lux (TechCrunch)

• Zoom unveils AI skills and agents for AI Companion (Health Tech Digital)

• What is Copilot Voice? How AI-powered commands can help (Microsoft)

• Jargonic from Aiola claims to best rivals at your business’s lingo (VentureBeat)

• Krisp receives 2 US patents on AI accent conversion (Krisp)

• Brain waves to spoken words: AI gives voice to people with paralysis (New Atlas)

• Voice AI assistant steals the show at 2025 Legalweek Conference (LexisNexis)

• AudioCodes leverages AI to help transform enterprise communication (TMCnet)

• How PolyAI is distancing itself in a crowded AI Voice market (TMCnet)

• Talkdesk’s agentic AI transforms global customer service (TMCnet)

• Micmonster.com reports major 2025 growth in AI TTS market (Markets Insider)

• Speaktor revolutionizes content creation with TTS technology (Markets Insider)

• AI-Media & AudioShake team up to transform live sports audio (GlobeNewswire)

• Krikey.ai launches AI-powered talking avatars with ElevenLabs (MarTech Cube)

• Decimal Point Analytics reshuffles leadership, gets new CTO (TechCircle)

• AI interviewing startup boosts hiring and candidate experience (HR Brew)

• Unlock insights in your contact centre conversations (Call Centre Helper)

• Smart Sound and Gateway Market 2025: AI audio innovation trends (OpenPR)

• Israeli AI-powered voice firm Auto raises $15M Series A (Tech in Asia)

• Ribbon raises $8M in funding (Finsmes)

• Sesame AI nears $1B valuation with $200M investment interest (WinBuzzer)

• Should your small business invest in an AI receptionist? (Tricky Android)

• Caantin unveils AI-driven communication solution (Tech in Africa)

---

Voice AI Podcast 🎙️

In case you missed the latest episode of Voice AI Podcast…

Engineering Corner 😎

• Qwen 2.5 Omni - The Most Multi-modal

• Boosting LLM for speech synthesis: An empirical study (Microsoft)

• People are poorly equipped to detect AI-powered voice clones (Nature)

• The minimalist’s guide to speech-to-text: Big wins with little data (HackerNoon)

• Introducing Dolphin: A multilingual ASR model optimized for eastern languages and dialects (MarkTechPost)

• Talking to the web: The rise of AI-powered voice navigation (TO THE NEW)

• Best TTS models: F5-TTS, Kokoro, SparkTTS, and Sesame CSM (DigitalOcean)

• Hailo demonstrates accelerated LLM-based speech recognition on the Raspberry Pi AI Hat (OurCrowd)