Yuma AI (YC W23)

When was the last time you bought something on your own website? Went through the full flow? The hurdles, the popups that don't close? The checkout button that disappears? Like, end to end, doing a real purchase, like a customer would? On this latest episode of CX After Hours, we got very lucky to host Zoe Kahn. She shared her 10% Fix: shop your own site. At least once a quarter, ideally every month, and every time you ship something new! Place an order, a real one. Check that the loyalty points actually landed. Read the emails that go out. Click those buttons that you added 2 years ago but haven't clicked since. In short, actively try to break things! A small change in one corner of the stack can quietly break something 3 steps downstream... If you do not find it, your customers will. And then, as always, it shows up in the support inbox. 😏 Which leads to one key point Zoe made in this episode: the support inbox is where the consequences of every decision in the business show up. It is the foundation of CX. Zoe previously led CX and retention at Chomps and Ghost, and now runs Inevitable Agency. Anya Kelly and I had a great time with her 🎙️ Full episode in the comments.

Today Yuma AI (YC W23) ships Voice. A real phone number, answered by AI, taking real actions in a merchant's backend while the customer is still on the line. Skip a delivery, fix an address, process a refund, confirmed out loud before anyone hangs up. This is the biggest thing we've shipped in 2026, and I want to be precise about why. It's not a phone bot bolted onto the side. It's the same agent merchants already run on email and chat, now picking up the phone. Same knowledge, same processes, same guardrails and hard limits. Nothing to retrain, nothing to configure twice. Anyone can put a voice model behind a number. Putting an agent behind it that already knows your policies, your catalogue and your customers, and can act on them mid-call, is a different thing entirely. On the phone there's no ticket queue to fall back on. The customer asks, and Yuma has to find the order, check the policy and make the change while they wait. That's the bar we built for. There's no draft to review, no delay to hide behind. The agent either resolves it or it doesn't. So the goal is full resolution, and Yuma goes for it. When a call lands outside the SOP or policy a merchant set, the handoff is built in: transfer the live call, collect an email and follow up, or route a ticket with the recording and transcript attached. You choose which. Every call gets recorded, every action on the record. 30+ languages, and it follows if the caller switches mid-sentence. Plugs into the helpdesk, commerce stack and telephony a merchant already runs. We've built a fictional coffee brand Daybreak Coffee so you can test it on real scenarios without anyone's data at risk: https://yuma.ai/voice-demo You can call it yourself, no phone number needed, straight from your browser. You can even test a guardrail by asking our Voice AI agent Emily for a refund, a new credit card, or an unrelated trivia question. She should decline and steer you back to Daybreak Coffee. Tell me if you get her off track! Huge thanks to the Yuma engineering team, who did the actual work here 💜 #ecommerce #customersupport #voiceAI #GTM #productlaunch

Back in December 2022, Yuma was the first company to put AI agents in production for ecommerce customer support. Almost 4 years later, we have had a pretty great run pushing that idea to its limits. But after 4 years building on top of a technology evolving this fast, something had to change... We call that change Yuma 2.0. One AI agent. One set of rules. Every customer. Every channel. Inbound and outbound. That is really the idea: a single, unified platform to run your entire CX operation. When we started Yuma, the challenge was getting an AI to reliably solve a customer support ticket. We had to do all kinds of gymnastics to get the models to behave, stay within the right boundaries and reliably take actions. Today, models are dramatically more capable. It has also never been easier to spin up a basic AI support agent. Give someone Claude or ChatGPT and they can probably get a convincing demo running in a few days. But a demo is one thing. Building something a company can trust to run at scale is another. More importantly, our customers are asking us for much more now. As I keep telling the team, bringing the most value to our customers is paramount. Today, CX teams are still spread across too many tools. One for support. One for voice. One for social. Another for outbound marketing. Different rules. Different context. Different levels of trust. I think that era is over. Our customers increasingly want to consolidate their customer experience around a single core system they can trust. So after almost 4 years of building Yuma, it was time to make another big leap. Part technological leap, part leap of faith, part founder instinct ;) Because I still deeply believe we are only seeing the very early glimpse of what AI will let us achieve. So with Yuma 2.0: • A single, unified set of rules ✅ • Inbound support across email, chat, social, reviews and voice ✅ • Outbound campaigns, from proactive service notifications to marketing and re-engagement ✅ • A deep understanding of every customer across every interaction ✅ • Public social engagement across Facebook and Instagram ✅ • Pre-sales conversations that help customers buy and increase conversion ✅ All powered by the same underlying AI agent. And that last detail matters: • Not a voice product bolted onto support. • Not a separate agent doing outreach. • Not 5 systems trying to understand the same customer. 👉 One agent, operating across the entire customer experience. We have baked almost 4 years of work, mistakes, customer feedback and learnings into this new version of Yuma. It is more powerful, more accurate, and lets our customers do things that simply were not possible when we started. I think this is where CX is going. And just like we did back in 2022, we want to be early in showing what that future can look like. That is Yuma 2.0. Come talk to us and see the mind-blowing things you can build today. Cheers to the next 4 years 🥂

my take is that all of the voice demo's we can get on the web doesn't prove it works i called a voice AI demo a few weeks ago from another company and it was really good, natural and it handled my use-case pretty well! (+ understood my terrible french accent) the thing is, a demo only proves the voice works. it doesn't prove the agent can actually do anything (ours included sorry Guillaume Luccisano) & on monday i heard a real call that we handled for a customer for a lost order, where we had to update the shipping address + reship the order in the backend, while the person was on the line. Holy cow it worked amazingly well! This is a true demonstration of what we can do on Voice, and the good news is that it's basically one-click away for our merchants. Link if you didn't read my message above: yuma.ai/voice-demo but if you want to see a real demo, on your customers, ping me :)

Mike Bair put it on our show better than we ever have: AI is a staffing strategy. 💡 10 years ago, he built a model that said: I have this much volume, I need this many people. Sounds simple, but that was the equation. It's still is, but now, AI is just another line in it. Which matters right now, because peak volume is forecastable and most teams already hold the inputs. Contact rate, order projections from the ops team or the marketing calendar. Naomi Oriol calls it demand planning: it's the same as inventory planning! So for Q4 this year, the question is not how many seasonal agents do we hire. It's: how much of this did we already know was coming? Which part of it actually needs a human? Hiring 25 agents for a 2-month crunch is what you do when you can't answer those questions. We built Yuma AI (YC W23) AI for the part you can forecast, so BFCM doesn't need a new hiring plan, every single year :)

Ask your CX team how painful it is to keep all policies, SOPs and processes up to date without breaking something else along the way. Most teams I talk to have the same setup. The return policy is in the bot, in the help center, in the post-purchase emails and in a handful of macros. When it changes, someone has to go through all of them and hope they didn't miss one. And because a lot of these rules depend on each other, like a warranty rule that references the return process, you usually find out about the one you missed when a customer gets the wrong answer. The same thing makes switching tools hard, and adding new ones just as hard. Every tool holds a piece of how your cs and cx works, so every change means rebuilding that piece somewhere else. With Yuma 2.0 we moved all of it into one place. Your rules, your policies and your customer context live in a Handbook, written in plain language. Everything around it can change. You can add a phone line, add SMS or try whatever new tool comes out next quarter, and they all follow the same rules, because the rules don't live in the tools anymore. It also stays easy to maintain as it grows. You tell Ask Yuma what changed, in your own words, from the dashboard or straight from Claude or ChatGPT. It updates the right pages, shows you the before and after, and runs your tests as simulated conversations before anything goes live, so if a change breaks another rule somewhere, you see it there and not in a customer reply. Your stack will keep changing. The rules your team spent years figuring out shouldn't have to move every time it does. The video shows one rule change going from a single sentence to the next customer reply. Link in the comments.