Events 🎤
Vapi and Deepgram kick off Web Summit with a voice AI mixer (Nov 9, Lisbon, Luma)
Decagon and Accel host a Voice & Audio Research Night (Nov 9, SF, Luma)
Cartesia launches Managed Agents in a live webinar, (Nov 10, Online, Luma)
ElevenLabs brings its Summit to New York(Nov 11, NYC, ElevenLabs)
VapiCon for two days at Fort Mason (Nov 11–12, SF, Vapi)
Telnyx takes Voice Connect to Dubai (Nov 12, Dubai, Luma)
Top Updates 💪
Salesforce agreed to acquire Cresta, a voice AI contact-center firm, in a deal reported at $2.7B to extend Agentforce into calls. TechTarget
Five9 is buying Level AI for $325M, paying cash to add AI-native contact-center tech. CMSWire
Dialpad and Ada are merging, an all-share deal to build an “agentic operating system” for customer work. CMSWire
PartnerOne acquired 3CLogic, folding its voice AI and contact-center layer into an enterprise portfolio. PR Newswire
ServiceNow open-sourced Apriel Voice 27B, a speech model that reasons and calls tools in one step. AI Weekly
Soniox launched STT v6, which feeds the agent’s own outgoing text to the recognizer to cancel its own voice. MarkTechPost
Phonic raised $27.5M and launched Phonic-1, a speech-to-speech model topping an industry benchmark. Slator
iZotope launched Photon, a local GPU app for AI stem separation and voice cleanup. Sonicstate
Krisp is a founding partner in Deepgram Marketplace, Deepgram’s reviewed directory of products for building voice agents. Deepgram
Apple is testing Audio Intelligence on Apple Watch, adding Live Rewind and Siri Recap in a watchOS 27.3 beta. The Verge
Microsoft filed a patent for AI that rewrites live speech, resyncing a speaker’s lips to the swapped words. Futurism
KT launched Voice Phishing Detection 3.0, Korea’s first multimodal detector reading a caller’s speech rate and tone on-device. BigGo
Scientific American explained spotting AI voices, noting jitter, shimmer and watermarks as detection signals. Scientific American
SoundHound expanded its Ultraviolette deal, rolling its voice co-pilot across the EV maker’s global lineup. Seeking Alpha
Taco Bell doubled down on drive-thru voice AI, now in 900+ restaurants with 1,500 targeted by year-end. Restaurant Business
AssemblyAI shipped TTS-1, a $15-per-million-character text-to-speech model aimed at live phone agents. Blockchain.News
Pypestream added voice to its contact center, routing AI-to-human handoffs on the same queues as its other channels. Yahoo
Otto launched a voice AI CRM, letting field reps update deals and draft follow-ups with a phone call. CXOToday
WellNet picked EhvaHealth’s voice AI to handle inbound health-plan provider calls end to end. PR Newswire
Aina unveiled a $499 AI smart ring, with a screen, joystick and mic to talk to and approve AI agents. TechCrunch
Smart Infinity launched PureVoice, on-network AI noise cancellation for calls with no app, toggled by dialing 8#. GadgetMatch
KrosAI raised $135K to connect voice agents to local phone networks across Africa and other emerging markets. WeeTracker
POLAR launched AISpeech in the UK and Ireland, distributing AI conferencing mics and voice-tracking cameras. LSi Online
Forbes argued voice agents are easy to demo but hard to ship, with production reliability the real engineering test. Forbes
Engineering Corner 😎
AWS documented Cartesia’s ultra-low-latency stack, deploying Sonic 3.6 on SageMaker for ~90ms speech generation. AWS
Nomono opened a spatial audio API, exposing speech separation and automatic 3D speaker tracking to developers. TechTimes
A paper introduced MURMUR, a long-form ASR inference system that cuts latency about 4x at single-pass accuracy. arXiv
A guide stress-tested 24/7 voice agents, laying out a 12-case availability checklist for production. dev.to