You can generate all the code in the world. But if your deployments take days, none of it matters. That's one of the sharpest observations from our webinar on how we doubled engineering productivity, and it gets at something most AI transformation conversations miss entirely. The teams moving fastest with AI aren't just the ones with the best tools. They're the ones who spent years building the culture and systems underneath. Fast deployments, shared technical standards and a unified way of working that means you solve a problem once and everyone benefits. AI didn't create that foundation, it just revealed how much it mattered. The full conversation is at the link in the replies.

Great AI support over email looks different than it does in chat. Here's how we built for that. In our recent Fin product update, Brian sits down with AI designer Burak to walk through the challenges of email support, and what we did to improve experiences. Customers expect long, well-formatted responses. PDFs need to be read. Spam, auto-replies, and CCs all need to be handled intelligently. And escalations that are easy to catch in chat can get buried in email. So that's what the Fin team built for. Fin now reads PDFs, formats responses for email, follows up automatically if it doesn't hear back, and handles noise with a new spam folder that shows you exactly why something was flagged. You can also write channel-specific guidance so Fin behaves differently over email than it does in chat, and test it all without leaving the product. Being excellently omnichannel means going deep on every channel. Catch up on all the latest Fin product updates at the link in the comments.

Most voice AI can answer a question. Fin Voice 2 can change your order, upgrade your shipping, and charge your card, all in the same call. Built on Apex Flash, our newest model purpose-built for latency-sensitive customer service, Fin Voice 2 doesn't just respond. It acts. In this clip, a customer needs their bike sooner than expected. Fin finds the expedited shipping option, confirms payment details, and updates the delivery, no transfer, callback, or waiting. The call ends with "happy biking." That's the bar. Learn more at the link in the replies

Many teams don't switch to a better AI Agent because they think they'll have to migrate their helpdesk to get it. They don't. Last week we announced Fin for Freshdesk and HubSpot. Your team stays in their tools, and your workflows don't change. You just get the world's best AI agent seamlessly integrated with them. Fin averages 76% resolution rates, with many teams hitting 90% once connected to their backend systems. Hundreds of teams are already running Fin on other helpdesks. Freshdesk and HubSpot are next. The decision about which AI Agent becomes the front door of your business should come down to performance, not which helpdesk it happens to be attached to. Open platforms win. Every support team deserves access to the best AI, regardless of the stack they're already on. Learn more at the links in the comments.

Learn more about Fin for Freshdesk here: fin.ai/freshdesk

Gizmo hurt his paw. His owner called and had an appointment booked in one minute. No hold time, form to fill out, or "press 1 for bookings." A natural conversation that got straight to the point, and a worried pet owner who hung up with exactly what they needed. That's Fin Voice 2. Built on Apex Flash, our newest model purpose-built for latency-sensitive customer service. It answers instantly, speaks naturally, and does more than answer questions. It resolves them. It takes action during calls. Verifying identities, processing refunds, changing subscriptions, booking appointments – all without transfers or hold times. For any business running on appointments, this is what great front-of-house looks like. Learn more at the link below.

Fin can now bring a teammate into a conversation, get a decision, and pick right back up where it left off. Brian sat down with Sze Wing to walk through exactly how it works. The use case: a VIP customer requesting a return outside the return window. Fin identifies the exception, pauses, notifies a teammate (in the inbox or directly in Slack), gets the approval, and completes the return. The customer never feels the seam. It's built for the moments where you want human judgment without handing over the whole conversation. High-value refunds, policy exceptions, cases where trust is still being built. The human makes the call. Fin handles everything else. Catch up on the latest Fin product updates at the link in the replies.

When AI handles the noise, the people who take on the sensitive, high-stakes work become harder to replace, not easier. That's how Danielle Constantine, CEO at MyHSA, sees it. The repetitive stuff disappears – what's left is the work that actually matters. And the people doing it are more valuable than ever. Danielle shared her thoughts on our recent webinar about how AI is reshaping careers in CX. The next one is on June 11 – find the 🔗 to register in the replies below.

Before going live with Fin, the Kalshi team tried to break it. At Fin Labs New York last week, Declan Ivory sat down with Shannon Magiera from Kalshi, and she walked through how that went. It started with a lesson from a previous platform. One engineer typed "talk to me like a pirate" – and it did. That was enough. If an AI Agent can be manipulated that easily, you can't trust it with your customers. They ran the same test on Fin, plus many more. It didn't budge. The nuance Declan added was perfect: one of Fin's customers actually *wants* their agent to talk like a pirate, and they can set that up intentionally. That's the difference between a guardrail and a limitation. Trust has to be the baseline. Everything else gets built on top of it. Watch the full session at the link in the comments.