Hamming AI

Voice AI's "demo to deployment" gap collapsed this week. Every layer of the stack shipped something material: TTS models claimed new leaderboards, agent platforms went self-serve, OpenAI and Google rolled out new realtime models, and NVIDIA shipped the first CPU built for agents. Here's what stood out: Cartesia dropped Sonic-3.5 and took #1 on Artificial Analysis's Speech Arena leaderboard in both global and open-weight rankings. PolyAI opened its Agentic Dialog Platform to every builder. The platform behind Marriott, FedEx, UniCredit, and Caesars is now free for the first two months, with Raven (trained on 1B+ enterprise conversations) as the default model. Rime shipped Coda, the next iteration of its TTS flagship. Dual-decoder architecture, on-prem deployable, and a serving stack (TIGERSTRIPE) that outperforms off-the-shelf implementations. Google I/O announced Gemini 3.5 Flash, Gemini Spark (24/7 agentic assistant), and voice features in Docs, Keep, and Gmail. Voice is now a default Workspace input, not a feature add-on. AssemblyAI released its Voice Agent API at $4.50/hr flat and announced Universal-3 Pro upgrades: 19% better WER on multilingual, 30% faster median latency. Modulate.ai released Velma Deepfake Detect at $0.25/hr, over 100x cheaper than competing solutions. #1 on Hugging Face's Speech Deepfake Arena. Deepfakes drove $1.6B in losses in 2025. NVIDIA delivered the first Vera CPUs (their first custom CPU, built for agentic workloads) to Anthropic, OpenAI, SpaceXAI, and Oracle. OCI plans to deploy hundreds of thousands this year. The pattern is clear: voice is no longer a single product category. It's a layered stack, and every layer just raised its bar.

We’re hiring a GTM Engineer at Hamming. Not an SDR. Not an AE. - You’ll do outbound, discovery, demos, closing, and expansion; basically, whatever gets customers successful. - Voice AI is a technical sale. You’ll be on screen shares with CTOs and engineering teams. Atypical backgrounds encouraged: ex-founders, lawyers, FDEs, chiefs of staff from small teams. Send us a one-pager, PDF, or Google Doc. Resume optional. 1. Why Hamming? 2. Why you? 3. What would you do in your first 30 days? 4. One strong opinion on voice AI.

"The proactive discovery is probably the number one most useful thing." - Jyoti Rani, AI Engineer at Synthpop Synthpop automates insurance verification end-to-end, cutting time-to-treatment and removing hours of manual work. Hamming helps make sure those voice agents keep working in production. We continuously test and QA real-world call scenarios before and after deployment, so QA doesn’t have to scale linearly with engineering. Huge credit to Jyoti and the Synthpop - Healthcare AI team for pushing the frontier of what agentic QA looks like in practice. Full case study 👇

Everyone talks about automating healthcare workflows. What’s talked about far less is the maybe even harder part: how do you rely on those systems once they’re live and handling real volume? Take insurance verification. You’re dealing with IVRs, fragmented payor logic, and a constant stream of edge cases that don’t behave nicely. At Synthpop, our voice agents run this workflow end-to-end across patients, clinics, and payors – cutting down time-to-treatment and removing a huge amount of manual effort. But getting the agent to work isn’t the finish line. Making sure it keeps working is. That’s where Hamming AI fits into our stack. We use it to continuously QA our voice systems – replaying and testing real-world call scenarios before and after deployment, without needing QA to scale with engineering. A lot of this has been pushed forward by our AI engineer, Jyoti, working closely with their team to shape what agentic QA actually looks like in practice. The shift is simple: Not just AI that works → AI that keeps working. Full case study: https://lnkd.in/d_kejEnm

We’re hiring a GTM Engineer at Hamming. Not an SDR. Not an AE. - You’ll do outbound, discovery, demos, closing, and expansion; basically, whatever gets customers successful. - Voice AI is a technical sale. You’ll be on screen shares with CTOs and engineering teams. Atypical backgrounds encouraged: ex-founders, lawyers, FDEs, chiefs of staff from small teams. Send us a one-pager, PDF, or Google Doc. Resume optional. 1. Why Hamming? 2. Why you? 3. What would you do in your first 30 days? 4. One strong opinion on voice AI.

Voice AI breaks differently than text AI: silent failures, exhausted QA, edge cases that only show up at scale. Three customer quotes from the last quarter: "The agent was saying 'I booked your appointment' but didn't actually book it. That's the kind of silent failure that destroys trust." - Josh Collin, CEO, Bland Labs "I find talking to the agents completely socially exhausting. The platform offers all of these different personas that the team is not able to replicate." - Blake Jones, AI Engineer, Basata "Not only can we have Hamming do it instead of us, we can have Hamming do it four or five times, all at once instead of having one person call and do it one time." - Tosh Toida, QA Lead, Mia Voice testing is engineering work. Hamming runs it for you.

Congrats to the Netomi Team on their $110M Series C led by Accenture Ventures, with Adobe Ventures and WndrCo. Agentic customer experience at Fortune 500 scale is one of the hardest deployment environments for AI agents. Netomi has been doing it since 2016. Delta, United, MetLife, Paramount, DraftKings, the NBA. Enterprise CX is full of AI demos that never make it to production. Netomi is the opposite. Excited to see what's next for the team.

Voice testing one customer at a time burns out your best engineer. "I find talking to the agents completely socially exhausting. The platform offers all of these different personas that the team is not able to replicate. We want to make sure that our diverse customer base is represented in the agents themselves." - Blake Jones, AI Engineer at Basata Healthcare voice agents need to handle every accent, every demographic, every emotional state. Hamming brings the persona library Basata's team can't build alone. Read the Basata case study. 👇

We're at LangChain Interrupt today. The whole event is about debugging and shipping reliable agents. That's our job at Hamming. If you're at Interrupt and shipping agents to production, let's chat.

Building voice agents is 70% testing. "It's like going from manual labor to using a tractor. You can prompt an agent in 30-45 minutes, but testing takes the next 2-3 hours. Building voice agents is 70% testing. Hamming makes that 70% manageable." Ahmad Rufai Yusuf, Forward Deployed Engineer at Bland Labs: Manual testing eats engineering days. Hamming runs them in parallel. Read the Bland Labs case study. 👇