Watch the full product update video here: piped.video/watch?v=tTrABO4j…

Tickets are where support work goes when it can’t be resolved in one reply, like a refund that needs finance or a bug that needs engineering. We recently shipped nine improvements to tickets. These updates are designed to improve the teammate, customer and cross functional teams experiences from the moment a teammate creates a ticket, to the way teams track that work and customers follow its progress. Three of the updates: → Fin autofills in the ticket title, description, and attributes from the conversation, so creating a ticket stops being a second round of data entry → A new lifecycle report shows how tickets move between teams, where they loop, and how long they sit with each one → You choose which status updates are worth an email, so customers stop tuning out the ones that matter Live for everyone today. Watch the full product updates video at the link in the comments.

Watch the full meetup conversation here: piped.video/6sR9KJFncrQ

There’s a reality that’s starting to take shape: for some teams, AI is already handling the majority of customer conversations. At our recent Fin x @AnthropicAI meetup, we shared how this is already happening in practice. Fin now resolves over 80% of support queries end to end. The technology is here, but adoption is still catching up. That progress is unlocking something bigger. Fin is moving beyond support toward a broader vision, where one agent can deliver a consistent, high-quality experience across the entire customer lifecycle, from day zero to year ten. Watch more below to hear how close we are to that future, and what it takes to get there.

Watch the full conversation about delivering high quality AI experiences here: piped.video/watch?v=-53JZG9-…

“Trust has been a multi-year journey.” At our recent Fin x @linear meetup, we explored what it actually takes to build trust in AI Agents, and why it starts with giving systems room to prove themselves, limiting their blast radius early on, and gradually expanding their role as customers see what they can reliably do without risk. Watch more below to hear how teams are measuring and earning that trust in practice.

Watch their full conversation here: piped.video/Zd0hvxUyLp4

What does it actually take to move from deploying AI to using it in ways that change how a business operates? That was the focus of our recent Fin x PostHog meetup in San Francisco. The conversation explored where companies are making real progress (and where things are still unsettled) – from the declining importance of traditional interfaces, to the realities of outcome-based pricing, and the challenge of driving true adoption inside organizations. A big thank you to Tim Glaser, Co-founder of PostHog, and Junan Pang, VP of Customer Success and Solutions at Fin, for a candid and thought-provoking discussion – and to everyone who joined us and brought sharp questions to the room. If you’d like to join a future Fin meetup, subscribe to our Luma calendar to see what’s next.

Watch their full conversation here: piped.video/6sR9KJFncrQ

“You only pay for Fin when it does the job.” At our recent Fin × @AnthropicAI meetup in Paris, this came up as a practical answer to a bigger question: how do you actually build trust in AI agents? For Jordan Neill, it starts with aligning incentives. Fin is designed to resolve customer problems end to end, and you only pay when it does. No usage-based ambiguity or paying for potential. Only outcomes. It’s a simple idea, but it changes the relationship entirely. Trust is built into how the product works. Watch the clip below. To join a future Fin meetup in your city, subscribe to our Luma calendar.

Watch the full conversation between Brian and Matthijs here: piped.video/-53JZG9-t20

Quality in AI isn’t just always a clean, single metric. It’s often something you feel when it works the way you expect, consistently. That’s what we explored at our Fin x @Linear meetup in London. @brian_donohue (Fin) and @matthijswolting (Linear) shared how teams are building for quality in practice, from combining AI scoring with human review, to creating visibility into what agents are doing and where they fall short. The throughline: trust is built by giving teams the tools to understand, test, and improve AI systems over time. Watch the highlights below. And if you want to watch the full conversation, see the link in the comments. If you’d like to join a future Fin meetup, subscribe to our Luma calendar to see what’s next.

Earlier this week in London, we partnered with @attio to explore how modern GTM teams are actually selling with AI, and where today’s approaches are still falling short. The conversation, led by @rati_zv (Fin) and @nicolasosharp (Attio), alongside @ciaran_nolan (Fin), focused on a growing mismatch: buyers are arriving faster, with more context and higher intent, but the systems they meet are still built for a slower, linear world. That gap set the stage for something new. We introduced Fin for Sales. It works like the ultimate salesperson: engaging prospects at just the right moment, guiding product discovery, suggesting tailored options, and qualifying in real time so no lead gets left behind. We also heard from Fin's @destraynor and Bethany Clark, and Attio's Will Jones, who shared how teams are starting to rethink the top of funnel, from static forms and handoffs to real-time, AI-led engagement. One theme that stood out was the idea of “lossless context.” Not just within a single interaction, but carried from the very first website visit all the way through to long-term customer relationships. As AI takes on more of the execution, the opportunity is to design systems that remember, adapt, and build on every interaction, so nothing gets lost as prospects become customers. From rethinking the role of the sales funnel to the emergence of new roles like GTM engineers, the discussion made one thing clear: this is more than simply adding AI to existing workflows, it’s redesigning them from the ground up. Learn more about Fin for Sales at the link in the replies 👇

Beyond the approval pipeline itself, our AI-authored backend code now has a revert rate of 0.53%, compared to 5.39% for human-authored. On the frontend, it’s 0.22% versus 2.00%. That means AI-authored code has a 10x lower revert rate.

The obvious: this process made us way faster. 5x faster to be precise. Median PR creation to approval time with AI: 14.6 minutes.

On the road to 2x, one of our biggest bottlenecks was PR approvals. Manual processes can't keep up. So, we created a PR review Agent. Today, over 19% of our PRs are auto-approved with no human reviewer in the loop. It might sound reckless but our data shows otherwise.

Watch their full conversation here: piped.video/6sR9KJFncrQ

"I don't know." That's the most powerful thing you can tell a customer. At our recent Fin × @AnthropicAI meetup in Paris, Ryan O'Holleran shared why honesty beats confidence when it comes to building trust — especially now, when AI can generate a plausible answer to almost anything. Fluency is easy. Knowing when not to answer is much harder. The best AI systems don't just get things right. They know when to pause, when to escalate, and when saying nothing is better than saying something wrong. Find the full conversation in the comments. And if you want to join a future Fin meetup in your city, subscribe to our Luma calendar 👇

New from @karenchurch – If Claude Code can do our jobs, what are we for? "AI commoditises execution. It does not commoditise judgment. And for research and data science teams, that's the difference between being relevant and being replaced." Read the full article here 👇 linkedin.com/pulse/claude-co…

You can watch the full meetup replay here: piped.video/-53JZG9-t20

What does ‘high quality’ actually mean in AI? At our recent Fin × @linear meetup in London, @matthijswolting (Engineering, Linear) put it simply: a lot of it is a vibe. You know it when an Agent does what you expect consistently, reliably, and without needing constant steering. The challenge is that feeling is hard to capture in numbers. @brian_donohue (VP Product, Fin) added the second layer: when it’s your product, quality also needs to be measured: resolution rates, real outcomes, whether the work actually gets done. That tension between what you can measure and what you can feel is where the best teams operate. Watch the clip below, and find the full conversation in the comments to go deeper.

You can watch the full conversation between Jordan and Ryan here: piped.video/6sR9KJFncrQ

“Last year it was: how do I make sure these AI models don’t hallucinate in a demo I’m showing internally? This year it’s: how do I make sure millions of customers are using my product, and I can sleep through the night?” That’s the moment AI is in right now, ensuring it can be trusted with real work. Because once AI is customer-facing, it becomes your brand. At our Fin × @AnthropicAI meetup in Paris, Jordan Neill (COO, Fin) and Ryan O'Holleran (GTM Leadership EMEA, Anthropic) kept coming back to the same idea: the technology is already there. What’s missing is the confidence to let it operate on your behalf – at scale. And that changes how you build. You design systems that know when not to answer. You draw hard lines around where humans stay in control. You build trust into the model, the product, and the way you bring it to market, not as a layer, but as the foundation. Watch their conversation below.

Watch their full conversation here: piped.video/AVVKO9cbRtw

Building an AI Agent for one company is straightforward. But making it work across every team, system, and dataset in any enterprise is where things get challenging. At our recent Fin x @Snowflake meetup in San Francisco, Jordan Neill (COO, Fin) and Baris Gultekin (VP of AI, Snowflake) broke down why context is the real bottleneck. As models improve, the challenge shifts to helping agents understand the nuances of each business (think data, language, and definitions) so they can operate reliably at scale. You can watch the full conversation at the link in the replies, and if you’d like to join a future Fin meetup, subscribe to our Luma calendar.

Complex queries are a small percentage of your support volume, but they take up most of your team’s time, and leave the biggest impression on customers. AI Agents have automated simple queries, but as soon as processes become complex and multi-step, things get harder to build, test, and maintain. You're typically told you need a forward-deployed engineer and weeks of engagement just to get started. And then more waiting every time something needs to change. We built Fin Procedures so your team owns this end to end, from creation to testing to monitoring. And we're not stopping there. We've been continuously shipping improvements (for example: AI-powered procedure review, failure reporting, version history, and email simulation) with more on the way. Since launching, Procedures has handled over 1.5M conversations across industries, with a ~28.9% proven improvement in CSAT. And now Procedures is available to every Fin customer.

Watch the full conversation here: piped.video/-53JZG9-t20

On March 31 we had a full room of CX and product leaders and builders in London for our Fin x @linear meetup on delivering high-quality AI experiences. Together with @brian_donohue (VP Product, Fin) and @matthijswolting (Engineering, Linear), we explored what it takes to move beyond demos and build AI systems people actually trust in production. One theme came up again and again: quality isn’t just a metric. It’s a combination of signals including resolution rates, CSAT, and the instinct you build from reviewing real conversations and outputs over time. We also got into the trade-offs that come with shipping AI at scale. From faster responses reducing trust, to new features temporarily impacting performance  and why oversight has to be built in from the start, not added later. Thanks to everyone who joined, asked thoughtful questions, and pushed the conversation forward. If you’d like to join a future Fin meetup, subscribe to our Luma calendar to see what’s coming next.

Watch the full conversation from Jordan and Baris here: piped.video/AVVKO9cbRtw

“It’s always easy to build demos and POCs, but the real world is messy.” That tension defined our recent Fin x Snowflake meetup in San Francisco, where we explored what it really takes to deploy Agentic AI in the enterprise. Jordan Neill (COO, Fin) and Baris Gultekin (VP of AI, @Snowflake) dug into the “iceberg” behind production systems – why demos are easy, but real-world deployment depends on shared context, strong evaluation loops, and architectures that can handle messy, unpredictable environments. As Agents move from internal tools to enterprise-wide systems, reliability (not raw capability) is what makes the difference. The shift is already underway. The technology is here, but adoption now depends on trust, feedback loops, and organizations rethinking how work actually gets done. Take a look at the highlights below, and if you’d like to join a future Fin meetup, subscribe to our Luma calendar.

Trust in AI isn't a feature. It's something you have to earn. That was the thread running through our conversation with @AnthropicAI at Fin Labs Paris, and it's harder than it sounds. Jordan Neill (COO, Fin) and Ryan O'Holleran (GTM Lead, EMEA, Anthropic) dug into the gap between what AI can do and what teams actually deploy. The technology isn't the bottleneck. Confidence is. When AI represents your brand, reliability, predictability, and knowing when not to answer matter as much as capability. Including the decision to say "I don't know." Want to be in the room for the next Fin meetup? Subscribe to our Luma calendar.

Watch the full conversation here: piped.video/7XBjliUOATQ

Incident response has traditionally been high-pressure, manual, and deeply human: a coordination challenge under stress. But with AI, a lot of that work is starting to shift. Searching through data, piecing together context, and extracting meaning in real time can now be supported by the tools teams use every day, reducing the pressure on the people responsible for getting things back on track. Watch the full conversation at the link in the comments, and if you’d like to join a future Fin meetup, subscribe to our Luma calendar.

Catch the full conversation here: piped.video/7XBjliUOATQ

The most important thing companies can do during an incident isn’t having all the answers. It’s showing up for their customers. Clear communication, acknowledging the issue, being present, and making it easy for customers to ask questions matters more than over-explaining or speculating. That’s what builds trust when things go wrong. Watch the full conversation at the link in the comments, and if you’d like to join a future Fin meetup, subscribe to our Luma calendar.

Catch the full conversation here: piped.video/7XBjliUOATQ

Why do we trust the companies we trust? Not because they never fail, but because of how they respond when they do. That was a key theme at our recent Fin x @incident_io meetup. You can watch the full conversation at the link in the replies. And, if you’d like to join a future Fin meetup, subscribe to our Luma calendar.

You can watch the full panel discussion from the meetup here: piped.video/x4gw-LxNk5s

"Being able to operate at the speed of thought is what I'm most excited about." At our recent Fin x @cartesia meetup in San Francisco, we asked attendees and speakers what’s exciting them most right now. This idea shone through from both our speakers onstage and the attendees in the room. It's clear: the pace of progress is compounding. What felt slow or experimental just months ago is now fast, usable, and closer to real-world deployment. From lowering the barrier to accessing technology to enabling more natural, real-time interactions, Voice AI is quickly moving from demo to default, and reshaping how people interact with software in the process. Take a look below and hear it in their own words.

What does it take to build an AI-native company in 2026? Next week in Paris, we’re opening the doors to Fin Labs for a full day with some of the city’s most ambitious founders for co-working and honest conversations about building with AI today. Plus, live demos showing what AI agents can actually do in production. On stage and in the room: ↳ Jordan Neill, COO, Fin ↳ Jean-Yves Poilleux, CTO & Co-founder, folk ↳ Louis Coppey, Partner, Point Nine Capital ↳ Andrew F., Solutions Engineer, Fin ↳ Faateh Dhillon, GTM, Dust ↳ Julien Frèche, Scaling Revenue folk Come to build, share what you’re working on, and connect with founders and investors shaping what’s next. Seats are limited, RSVP via Luma below 👇

We recently partnered with incident.io in London to dig into a topic every team eventually faces: incident response. The conversation, led by Declan Ivory (Fin), Lawrence Jones (@incident_io), and Eric Fitzgerald (Fin), focused on what actually separates high-performing teams when things go wrong. One key theme was that incidents are operational challenges, but they’re also trust-defining moments. Your customers won't remember the outage as much as they'll remember how you showed up. From declaring incidents early to using AI to take some of the pressure off in the moment, the discussion revolved around the idea that great incident response comes down to clarity, coordination, and consistent communication. If you’d like to join a future Fin meetup in your city, subscribe to our Luma calendar 👇

Last week in San Francisco, we partnered with Snowflake for a candid conversation on what it really takes to deploy agentic AI in the enterprise. Onstage, Jordan Neill (COO, @intercom) and Baris Gultekin (VP of AI, @Snowflake) compared notes on moving from impressive demos to production systems that actually work reliably, securely, and at scale. One theme came up again and again: the iceberg. It’s easy to “vibe code” a prototype in a week. It’s much harder to build the semantic layers, eval systems, guardrails, and feedback loops that make an agent trustworthy in the real world. They also dug into: ↳ Why evals are necessary but never enough without real-world signal ↳ How Snowflake brings AI to the data (not data to the AI) ↳ What it takes to make multi-step workflows (like refunds) actually reliable ↳ Why enterprise adoption is more about habit change than model capability Thank you to Baris and the Snowflake team for the partnership, and to everyone who packed the room, asked thoughtful questions, and stayed late to continue the discussion. If you’d like to join a future Fin meetup in your city, subscribe to our Luma calendar to see what’s next.

Watch the full meetup here: piped.video/7XBjliUOATQ

"The companies customers trust most aren't the ones that never fail. They're the ones that respond well when they do." That framing opened our recent Fin × @incident_io meetup in London. Declan Ivory (VP of Customer Support at Intercom, building Fin) and Lawrence Jones (Founding Engineer at incident.io) explored what excellent incident leadership actually looks like—from the moment something breaks to the post-mortem that rebuilds trust. A few themes stood out from the conversation: 𝗗𝗲𝗰𝗹𝗮𝗿𝗲 𝗲𝗮𝗿𝗹𝘆, 𝗱𝗲𝗰𝗹𝗮𝗿𝗲 𝗼𝗳𝘁𝗲𝗻. Most teams don't declare incidents frequently enough—and the incentives behind that are harmful. Low-friction declaration builds the muscle you need for serious incidents. The question isn't whether to start an incident that turns out not to be one. It's whether you can afford not to. 𝗔𝗜 𝗿𝗲𝗱𝘂𝗰𝗲𝘀 𝗰𝗼𝗴𝗻𝗶𝘁𝗶𝘃𝗲 𝗹𝗼𝗮𝗱, 𝗻𝗼𝘁 𝗵𝗲𝗮𝗱𝗰𝗼𝘂𝗻𝘁. The biggest opportunity for AI in incident response is absorbing the coordination overhead—summarizing root cause, tracking remediation, handling the customer surge—so humans can think clearly when the judgment calls matter most. Watch the highlights below, and follow the link below to watch the full discussion.

Watch the full discussion here: piped.video/x4gw-LxNk5s

Voice AI has crossed the uncanny valley, and now the real work begins. At our recent Fin × @cartesia meetup in San Francisco, Peter Bar (leading Voice AI at @intercom, building Fin Voice) and Iz Shalom (Head of Product at Cartesia) unpacked what it actually takes to scale voice agents from prototypes to millions of real customer calls. A few moments from the conversation stood out: → 𝗪𝗲’𝘃𝗲 𝗵𝗶𝘁 𝗮𝗻 𝗶𝗻𝗳𝗹𝗲𝗰𝘁𝗶𝗼𝗻 𝗽𝗼𝗶𝗻𝘁: “The beautiful thing with an exponential curve is no matter where you are, it’s still exponential.” Voice quality has steadily improved for two years, and we’re still early. Customers are no longer immediately demanding a human. They’re engaging. → 𝗬𝗼𝘂’𝗿𝗲 𝗷𝘂𝗱𝗴𝗲𝗱 𝗯𝘆 𝘆𝗼𝘂𝗿 𝘄𝗼𝗿𝘀𝘁 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻: In creative AI, you’re evaluated on your best output. In voice agents, you’re evaluated on the one awkward pause, the strange whisper, the off answer. Reliability under pressure is what makes or breaks production systems. → 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗶𝘀 𝗺𝗼𝗿𝗲 𝘁𝗵𝗮𝗻 𝗹𝗮𝘁𝗲𝗻𝗰𝘆: Early focus was naturalness and speed. Then customers asked the harder question: how does this integrate into my workflows? Voice AI has to solve real problems, hand off cleanly, and fit into existing operations, not just sound good in a demo. Watch the highlights in the video below. And if you want to watch the full discussion, the complete meetup recording is linked in the replies 👇

We're launching v2 of the AI Startup Pack by @fin_ai, a new toolkit built to help startups move faster with $500,000+ in AI credits. Through the pack, startups can get discounts on AI products from @fin_ai, @elevenlabs, @cloudflare, @incident_io,@attio, and more. If you’re building an AI-powered startup, check it out.

Sze shares how with Procedures, Fin becomes a collaborative teammate. Your human agents and Fin work together to answer your customer queries in whichever tool you’re using – Salesforce, Zendesk or @Intercom. Fin works seamlessly with your teammates to leave notes and resolve your customer conversations with speed and accuracy. Watch the full Procedures and Simulations video for more – link in the replies.

What types of customer conversations can Fin solve? Now, Fin can handle complex queries and edge cases like refunds, data deletion and finding deposits. Luis explains how with Procedures, Fin solves thousands of these complex and sensitive scenarios which used to need bespoke attention from your human support team. Explore more about Procedures and Simulations in Fin - video linked in the comments.

We built Simulations in Fin to give support teams confidence in using automation at scale. Here, Sze explains how to run automated AI tests to verify that Fin behaves as you intend in different customer scenarios. Watch the full video, linked in the replies below.

In case you missed it, last week we announced big updates to Procedures and Simulations in Fin. These features make AI automation for complex queries more powerful, and simpler for support teams to manage themselves. Watch the full video of what’s new in Procedures & Simulations from our CPO Paul Adams and the team.

In less than 1 minute, Luis explains how Procedures work and the important sweet spot. Build your AI agent for even the most complex queries using natural language, all whilst maintaining precision and control at the moments that matter most. And it scales across tens of thousands of conversations. For more, watch the full Procedures and Simulations: fin.ai/procedures#product-up…

Legacy players don't have to become casualties. We’re the proof. It takes hard work but anyone can become an AI company if they’re prepared to take big risks. In our latest Fin Weekly, we cover: → The moment: Bloomberg's Odd Lots names us a pivot success story. → The product: 12 major updates making Fin capable of your most complex workflows. → The frontier: Why voice AI may have finally crossed the uncanny valley. Read the full edition below:

The phone isn’t going away, but the way we build for it 𝗶𝘀 changing. This was the focus of our recent Fin x @cartesia meetup at @intercom HQ in San Francisco. Onstage, Peter Bar, Product Lead for Fin Voice at Intercom, and Israel Shalom, Head of Product at Cartesia, talked through the clear shift: the era of scripted voice workflows is over. What replaces it is far more powerful, and also far more demanding. Shipping Voice AI Agents in production requires accounting for resolution rate, response speed, interruption handling, background noise, and answer quality that holds up across millions of conversations. As Peter put it, you’re not judged on your best generation. You’re judged on your worst one. Thanks to Israel and the Cartesia team for the partnership, and to everyone who brought thoughtful questions and hard-earned perspective to the room. If you’d like to join a future Fin meetup, subscribe to our Luma calendar to see what’s next.

And you can watch the full meetup recording here: piped.video/watch?v=tqT45PYQ…

"It's extremely important to be super focused on the customer, to be customer obsessed. It's very easy to build very fast. The one problem that we need to focus on is what to build." That was one of the clearest takeaways from our Fin × @productboard meetup in San Francisco. In a room full of product leaders navigating an AI-first world, the reflections echoed what Archana Agrawal (@intercom, building Fin) and Hubert Palan (Productboard) discussed on stage: AI is creating incredible opportunities to build faster and push products further, but it doesn’t remove the need for judgment. Across the conversations, a few themes kept coming up: → Stay relentlessly customer-obsessed, even as velocity increases → Take the speed of innovation and translate it into something humans can actually digest → Get clear on the exact pain point before you start building: more user interviews, more validation, fewer assumptions AI is here to augment product teams. The challenge (and the opportunity) is using that power to deliver real value. Watch the highlights below, and subscribe to the Fin Luma calendar to join future meetups in your city.

Full meetup recording here: piped.video/watch?v=tqT45PYQ…

“Agentic coding is going to help you build faster; but are you building the right thing?” That question framed our recent Fin × @productboard meetup in San Francisco. Archana Agrawal (President, Intercom, building Fin) and Hubert Palan (Founder & CEO, Productboard) explored what product craft looks like when AI removes speed as the bottleneck. In a world where PRDs, prototypes, and features are easier than ever to produce, staying focused on customer problems becomes even more important to build products that add real value. A few themes stood out from the conversation: 𝗣𝗿𝗼𝗱𝘂𝗰𝘁 𝗰𝗿𝗮𝗳𝘁 𝗶𝘀 𝘀𝘁𝗶𝗹𝗹 𝗰𝘂𝘀𝘁𝗼𝗺𝗲𝗿 𝗳𝗶𝗿𝘀𝘁. Great products come from deeply understanding what people are struggling with, and translating that into decisions that create real value. AI can help synthesize information, but responsibility still sits with product leaders to make the decisions and own the outcomes. 𝗦𝗽𝗲𝗲𝗱 𝗶𝘀 𝗻𝗼 𝗹𝗼𝗻𝗴𝗲𝗿 𝘁𝗵𝗲 𝗲𝗱𝗴𝗲. When everyone can build fast, differentiation comes from focus, taste, and deciding what not to build. AI can accelerate execution, but velocity only matters if you’re moving in the right direction. Watch the highlights below, and follow the link below to watch the full discussion.

“AI can help you build faster, but it can’t decide what deserves to exist.” That idea was key at our meetup last week in San Francisco, where we partnered with @productboard for a candid conversation on product craft in the age of AI. Our President, Archana Agrawal, sat down with Hubert Palan, Founder & CEO of Productboard, to explore how teams should think about building products to solve real customer problems. The throughline was clear: In a world where AI can draft PRDs, summarize feedback, and automate execution, speed is no longer the bottleneck. As AI makes it easier to build almost anything, product craft becomes the discipline of choosing what to build with focus and intention, and saying no to everything else. Thanks to Archana and Hubert for such an honest, insightful discussion, and to everyone who joined and pushed the conversation further with thoughtful questions. If you’d like to join a future Fin meetup, you can subscribe to our Luma calendar to see what’s next 👇

You can watch the full Fin Product Updates video here: piped.video/2DoTrykJ3yo

If your team serves other businesses, support conversations often creep into Slack. So we've launched Fin over @SlackHQ, where Fin can answer customer questions directly in Slack. It works exactly as you’d expect: replies stay threaded, context remains intact, emojis work, and all conversations are visible and tracked in your inbox. For more of what’s new, watch the full Fin Product Updates – December 2025 video ↓

To see what else is new with Insights, you can watch the full December Product Updates video here: piped.video/2DoTrykJ3yo

As support volume grows, it's difficult to track what's driving performance. Fin already groups conversations into Topics and Subtopics, providing clear visibility over customer conversations. Now Topic Trends automatically surfaces the most interesting changes, like spikes or drops in conversation volume, dips in CX Score, or shifts in how well Fin is resolving questions. Instead of digging through reports, you get a clear, focused list of what deserves attention, backed by real customer conversations and delivered automatically. This clip shows Topic Trends in action. For more, watch the full Fin Product Updates – December 2025 video ↓

For more on what's new with CX Score, you can watch the full December Product Updates video here: piped.video/2DoTrykJ3yo

“It's easy and fast to get to a thin AI product. But if this is something that’s going in front of your customers, or evaluating how good Fin and your team is doing, it's got to be something you can trust.” Hear Brian and Christine explain why trust in AI metrics is uniquely high stakes. And why we’ve spent months putting CX Score through deep statistical evaluation and human review of thousands of real conversations before shipping it.

Watch the full December Product Updates video here: piped.video/2DoTrykJ3yo

Better performance comes from continued investment. That applies to most things in life, including your AI Agent. And it's why we've created the Fin Flywheel: the continuous improvement system behind Fin. It's how you train, test, deploy, and analyze performance so Fin improves the more you use it. Brian shows how the Fin Flywheel works in practice and how we organize every feature we build around it. To see what else is new, you can watch the full Fin Product Updates – December 2025 video ↓

Watch the full Fin Product Updates – December 2025 video here: piped.video/2DoTrykJ3yo

CSAT only measures ~8% of your conversations. That's never been good enough. It's just been your only option, until now. Stage 1: In 2025 we launched CX Score, an AI-powered metric that measures customer experience across every conversation, no surveys required. Stage 2: We've just launched CX Score Reasons, which gives you visibility into what's driving each score. Whether that's high user effort, low answer quality, or negative product feedback, you can explore these reasons at the conversation level, roll them up into reports, track trends over time, and layer them with topics for deeper context for your business. Now, CS leaders can quickly know what's working, what's broken, and what to prioritize to drive continuous CX improvements. For more of what’s new, watch the full Fin Product Updates – December 2025 video ↓

“What’s the stuff that wasn’t obvious and is hard to do? What’s the stuff most people don’t really realize, but that we feel the pain of?” The panelists set out to answer these questions during our Fin x @WeAreLegora x @InsidePersonio meetup at @SlushHQ in Helsinki, where the panelists reflected on what they’ve learned after three years of building in the AI era. A few themes shone through: 𝗔𝗜 𝗮𝗱𝗼𝗽𝘁𝗶𝗼𝗻 𝗶𝘀 𝗹𝗲𝘀𝘀 𝗮𝗯𝗼𝘂𝘁 𝘁𝗼𝗼𝗹𝘀 𝗮𝗻𝗱 𝗺𝗼𝗿𝗲 𝗮𝗯𝗼𝘂𝘁 𝘁𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻. One speaker put it plainly: “There’s loads of AI tools available, that’s not the bottleneck. AI should not be considered as these tools and solutions, but rather the transformation in the organization.” Time and time again, the conversation returned to the importance of helping customers manage change, define workflows, and get set up for success. 𝗔𝗜 𝗵𝗮𝘀 𝗰𝗵𝗮𝗻𝗴𝗲𝗱 𝘄𝗵𝗮𝘁’𝘀 𝗽𝗼𝘀𝘀𝗶𝗯𝗹𝗲, 𝗯𝘂𝘁 𝗻𝗼𝘁 𝘄𝗵𝗮𝘁’𝘀 𝘃𝗮𝗹𝘂𝗮𝗯𝗹𝗲. The panel talked about how AI has accelerated iteration cycles. One speaker took the point further, “You have an idea. The next day you’re already able to implement it, test it, and see if it works.” While AI has sped up product development, it has also raised the bar for judgment about what’s actually worth building. 𝗧𝗵𝗲 𝗹𝗶𝗺𝗶𝘁𝘀 𝗮𝗿𝗲 𝘀𝘁𝗶𝗹𝗹 𝗿𝗲𝗮𝗹, 𝗮𝗻𝗱 𝘁𝗵𝗲𝘆 𝗺𝗮𝘁𝘁𝗲𝗿. The discussion also touched on what AI can’t do (yet). As one speaker said, “We can’t think that we have everything available right now. We should always aim for something which is not doable right now, but might be doable in two years.” Leading in this era means walking the line between ambition and realism. Watch the highlights below, and subscribe to the Fin Luma calendar to join future meetups in your city.

“The real value is in being authentic and not worrying too much about getting everything perfect, just getting the message across.” That idea sat at the centre of our Fin x Clay meetup in London, where Paul Adams and Varun Anand explored how GTM and marketing are being reshaped in the AI era. A few themes stood out: 𝗔𝘂𝘁𝗵𝗲𝗻𝘁𝗶𝗰𝗶𝘁𝘆 𝗶𝘀 𝗯𝗲𝗰𝗼𝗺𝗶𝗻𝗴 𝗮 𝗰𝗼𝗺𝗽𝗲𝘁𝗶𝘁𝗶𝘃𝗲 𝗮𝗱𝘃𝗮𝗻𝘁𝗮𝗴𝗲. In a world full of AI-generated sameness, being more human and less polished makes brands more memorable. This resonated with attendees, including one who said “Everyone being a bit more human, a bit more real… those unique human insights are the things that are gonna make you stand out.” 𝗚𝗿𝗲𝗮𝘁 𝗺𝗮𝗿𝗸𝗲𝘁𝗲𝗿𝘀 𝗰𝗮𝗻 𝗰𝗼𝗺𝗲 𝗳𝗿𝗼𝗺 𝗮𝗻𝘆 𝗱𝗶𝘀𝗰𝗶𝗽𝗹𝗶𝗻𝗲. One attendee summed it up simply: “A lot of marketers don’t start out traditionally in marketing, but they start with a passion and they adopt new tools and AI.” This echoed the panel conversation about hiring for curiosity, taste, and the ability to ship, rather than for conventional marketing backgrounds. 𝗥𝗼𝗹𝗲𝘀 𝗮𝗿𝗲 𝗲𝘅𝗽𝗮𝗻𝗱𝗶𝗻𝗴, 𝗻𝗼𝘁 𝗻𝗮𝗿𝗿𝗼𝘄𝗶𝗻𝗴. The discussion pointed toward fewer specialists and more full-stack owners. One speaker noted that teams are finding “new ways to explore existing channels and do the job of what used to be three people.” Gone are the days where you needed a perfectly polished marketing campaign. To cut through the noise today, you need strong storytelling, authenticity, and teams willing to experiment and hit publish before everything is fully polished. Watch the highlights below, and subscribe to the Fin Luma calendar to join future meetups in your city.

16 meetups. 10 cities. 35 AI leaders and innovators. 1600+ attendees. The goal across our Fin meetups this year was simple: help people understand how AI is transforming the way we work, reshaping organizations and team structures, and redefining how products are built, sold, and priced. Leaders from companies like @linear, @Lovable, @attio, @clay, and more shared valuable insights into how they’re navigating this shift, what’s working and what isn’t. A few big themes kept coming up across these conversations: 𝗣𝗮𝗰𝗲 𝗮𝗻𝗱 𝗽𝗿𝗼𝗴𝗿𝗲𝘀𝘀 𝗺𝗮𝘁𝘁𝗲𝗿 𝗺𝗼𝗿𝗲 𝘁𝗵𝗮𝗻 𝗽𝗲𝗿𝗳𝗲𝗰𝘁𝗶𝗼𝗻: In a world where surface features can be copied quickly, speed, persistence, and product momentum are what separate long-term winners from short-lived copycats. 𝗙𝗮𝗹𝗹 𝗶𝗻 𝗹𝗼𝘃𝗲 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗽𝗿𝗼𝗯𝗹𝗲𝗺 (𝗻𝗼𝘁 𝘁𝗵𝗲 𝘁𝗲𝗰𝗵): The best teams are obsessing over building trustworthy, reliable AI that solves real problems and delivers outcomes. 𝗥𝗼𝗹𝗲𝘀 𝗮𝗿𝗲 𝗰𝗼𝗻𝘃𝗲𝗿𝗴𝗶𝗻𝗴: The shape of teams is shifting toward more hybrid profiles, people who care about the full experience and are happy to do what it takes, instead of staying strictly in their lane. Watch the video below to catch some memorable moments and standout ideas from our speakers to spark your own for next year. How has AI changed the way you work this year? What will you carry into next year?  Let us know in the comments below. Thank you to everyone who joined us in person (or caught the recaps and replays)! Your curiosity, energy, and participation helped enrich the conversations we're having on AI. We can't wait to bring you an exciting schedule of new meetups with forward-thinking leaders in 2026. Stay tuned.

“Everyone out there has demos… you can whip up a demo super fast. But to actually make it production ready, customer ready, is a huge difference.” At our Fin x @WeAreLegora x @InsidePersonio meetup at Slush in Helsinki, Brian Donohue (@intercom) sat down with Max Junestrand (Legora) and Hanno Renner (Personio) to explore what it really takes to lead change in the age of AI. Here are some key takeaways from their conversation: 𝗔𝗜 𝗼𝗻𝗹𝘆 𝗺𝗮𝘁𝘁𝗲𝗿𝘀 𝘄𝗵𝗲𝗻 𝗶𝘁 𝘀𝗼𝗹𝘃𝗲𝘀 𝗿𝗲𝗮𝗹 𝗰𝘂𝘀𝘁𝗼𝗺𝗲𝗿 𝗽𝗿𝗼𝗯𝗹𝗲𝗺𝘀. The panel emphasized that AI should create value by improving real workflows, not by adding features for their own sake. As Hanno put it, AI brings value where it solves an existing customer need in a more powerful way. 𝗔𝗜 𝘁𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 𝗿𝗲𝗾𝘂𝗶𝗿𝗲𝘀 𝗵𝘂𝗺𝗮𝗻 𝘁𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻. Buying AI tools is just step one. “Transformation” (a word Brian joked is painfully cliché but accurate) requires handholding. Customers still need help managing change inside their orgs, defining workflows, and teams need onboarding, setup, and troubleshooting support. Services and training now matter more than ever, and often are a deciding factor for AI transformation. 𝗡𝗲𝘄 𝗿𝗼𝗹𝗲𝘀 𝗮𝗿𝗲 𝗲𝗺𝗲𝗿𝗴𝗶𝗻𝗴 𝗮𝘀 𝘁𝗲𝗮𝗺𝘀 𝗮𝗱𝗮𝗽𝘁. AI is reshaping org design and creating roles at the intersection of domain expertise and AI systems. Max introduced Legora’s new role of Legal Engineers, who behave almost like consultants: interviewing partners, documenting deal workflows, and turning internal processes into structured logic. Watch the highlights below, and subscribe to the Fin Luma calendar to join future meetups in your city.

“You have to focus on the problem. As building gets easier, it's about the strategy and discovery that really makes your product stand out.” That attendee reflection summed up a core theme from our London meetup with @attio and @meetgranola, where @DesTraynor, @nicolasosharp (Nick), and @VasZhovner explored what product strategy really means in the AI era. A few insights from the panel resonated throughout the room: 𝗣𝗹𝗮𝗻 𝗳𝗼𝗿 𝘂𝗻𝗰𝗲𝗿𝘁𝗮𝗶𝗻𝘁𝘆. All three speakers agreed that AI has made long-range plans harder and uncertainty the norm. Attendees felt that too. Vas put it plainly: “It is still very early days for AI and with technology evolving so fast, it is very hard to predict how our future is going to look like.” 𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗮𝘁 𝘁𝗵𝗲 𝗳𝗿𝗼𝗻𝘁𝗶𝗲𝗿 𝗿𝗲𝗾𝘂𝗶𝗿𝗲𝘀 𝗲𝘅𝗽𝗹𝗼𝗿𝗮𝘁𝗶𝗼𝗻 𝗮𝗻𝗱 𝗿𝗲𝘀𝘁𝗿𝗮𝗶𝗻𝘁. The panel talked about separating fast, messy experimentation from mission-critical systems that require evaluation and discipline. A good AI product strategy is knowing which parts of your product belong in which mode. As Nick put it, “We are in truly uncharted territory, and we’re building things at the frontier of technology.” 𝗔𝗜 𝗰𝗵𝗮𝗻𝗴𝗲𝘀 𝗵𝗼𝘄 𝘆𝗼𝘂 𝗯𝘂𝗶𝗹𝗱 (𝗮𝗻𝗱 𝘄𝗵𝗼 𝘆𝗼𝘂 𝗵𝗶𝗿𝗲), 𝗻𝗼𝘁 𝗷𝘂𝘀𝘁 𝘄𝗵𝗮𝘁 𝘆𝗼𝘂 𝗯𝘂𝗶𝗹𝗱. Des highlighted that teams must rethink processes, roles, and expectations: “You probably need to change more than you realize inside your company on all fronts. You need to change how you build, what you build, why you build, who you hire.” 𝗩𝗲𝗹𝗼𝗰𝗶𝘁𝘆 𝗺𝗮𝘁𝘁𝗲𝗿𝘀, 𝗯𝘂𝘁 𝗼𝗻𝗹𝘆 𝘄𝗵𝗲𝗻 𝗶𝘁 𝘀𝗲𝗿𝘃𝗲𝘀 𝘀𝗼𝗹𝘃𝗶𝗻𝗴 𝗿𝗲𝗮𝗹 𝗽𝗿𝗼𝗯𝗹𝗲𝗺𝘀. The panel discussed how AI has shortened iteration loops and amplified the need for good judgment. Attendees echoed that duality. One said her favorite part of working with AI is the velocity, the speed in which you can fail, and learn. Watch the highlights below, and subscribe to the Fin Luma calendar to join future meetups in your city.

“Not everybody that cloned you in five minutes can also keep up with you.” That line from an attendee at our Lisbon meetup with @glean and @heyjasperai captured a core theme of the panel. AI has lowered the cost of imitation, but it has dramatically raised the bar for sustained execution. A few insights came through (both from the speakers onstage and from the builders in the audience): 𝗘𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻 𝗶𝘀 𝘁𝗵𝗲 𝗺𝗼𝗮𝘁. The panelists were clear that velocity, iteration, and deep integration work are the real differentiators in the AI era. It’s not the demo that matters, it’s whether the product works in real workflows. And, as one attendee put it, “The biggest differentiator is execution and consistency.” 𝗔𝗜 𝗶𝘀 𝗿𝗲𝘀𝗵𝗮𝗽𝗶𝗻𝗴 𝘄𝗵𝗮𝘁 𝘁𝗲𝗮𝗺𝘀 𝘀𝗽𝗲𝗻𝗱 𝘁𝗵𝗲𝗶𝗿 𝘁𝗶𝗺𝗲 𝗼𝗻. The panel explored how AI is helping automate repetitive work and freeing teams to focus on creativity, strategy, and orchestration. Voices from the room echoed that shift in real terms: “We’re now able to do more exciting work because a lot of the boring work is gonna get automated.” 𝗦𝗽𝗲𝗲𝗱 𝗶𝘀 𝗻𝗼𝘄 𝘁𝗵𝗲 𝗼𝗽𝗲𝗿𝗮𝘁𝗶𝗻𝗴 𝗰𝗼𝗻𝘀𝘁𝗿𝗮𝗶𝗻𝘁. The panel emphasized that what used to take years now takes months or weeks, and that companies must adapt their rhythms accordingly. Attendees reflected the same reality, “Building it is no longer three years or four years. It can be a few months or a few weeks.” Across both the panel and the hallway conversations, the message was consistent: in the AI era, speed, outcomes, and persistent execution beat novelty. The teams that keep improving are the ones the market will not be able to catch. Watch the attendee reactions below, and subscribe to the Fin Luma calendar to join future meetups in your city.

“The best words are the simple words. The best words are the words people understand.” That line from Paul Adams captured a key idea from our recent meetup with @clay, where he and Varun Anand shared how they think about GTM and marketing in an era when it’s hard to capture attention, people are tired of corporate sludge and AI slop, and taste matters more than ever. Some profound themes stood out: 𝗦𝗶𝗺𝗽𝗹𝗶𝗰𝗶𝘁𝘆 𝗶𝘀 𝗮 𝗰𝗼𝗺𝗽𝗲𝘁𝗶𝘁𝗶𝘃𝗲 𝗮𝗱𝘃𝗮𝗻𝘁𝗮𝗴𝗲. Clear language builds trust, especially when the product or the tech behind it is complex. As Paul put it, “Is that how you'd explain it to someone in the room? No? Then why are you using complex words externally?” 𝗚𝗿𝗲𝗮𝘁 𝗺𝗮𝗿𝗸𝗲𝘁𝗲𝗿𝘀 𝗰𝗮𝗻 𝗰𝗼𝗺𝗲 𝗳𝗿𝗼𝗺 𝗮𝗻𝘆 𝗳𝘂𝗻𝗰𝘁𝗶𝗼𝗻. The people who generate the most resonant ideas are the ones who’re deeply curious and passionate, and who already create and ship things. Varun stated that marketing is a skill set that you can find in really any function. 𝗔𝘂𝘁𝗵𝗲𝗻𝘁𝗶𝗰𝗶𝘁𝘆 𝗶𝘀 𝗼𝘂𝘁𝗽𝗲𝗿𝗳𝗼𝗿𝗺𝗶𝗻𝗴 𝗽𝗼𝗹𝗶𝘀𝗵. As Varun put it, consumer marketers “have to be really cutting edge” because they’re competing with TikTok and Instagram feeds, and B2B buyers are in those same feeds too. Modern GTM needs more real moments, and a willingness to show work that isn’t over-produced. 𝗧𝗮𝘀𝘁𝗲 𝗮𝗻𝗱 𝗷𝘂𝗱𝗴𝗺𝗲𝗻𝘁 𝗮𝗿𝗲 𝗻𝗼𝘁 𝗼𝗽𝘁𝗶𝗼𝗻𝗮𝗹. AI accelerates content creation, but it also makes the output look and sound the same. Modern teams need people with taste, people who know when something is actually good, not just correct. That last 20%, the judgment to elevate a piece of work, is what helps cut through the noise. Watch the highlights below, and subscribe to the Fin Luma calendar to join future meetups in your city.

“The best thing you can do is to learn as fast as possible. You don’t know what a great experience is going to look like until you actually build it and show it to users.” - Vas Zhovner That line set the pace at our London meetup with @attio and @meetgranola, where @destraynor, @nicolasosharp, and @VasZhovner unpacked what it really means to lead product strategy in the AI era. Some key takeaways stood out: 𝗥𝗼𝗹𝗲𝘀 𝗮𝗿𝗲 𝗰𝗼𝗻𝘃𝗲𝗿𝗴𝗶𝗻𝗴: The shape of teams is shifting toward more hybrid profiles – people who care about the full experience and are happy to do what it takes. 𝗥𝗼𝗮𝗱𝗺𝗮𝗽𝘀 𝗮𝗿𝗲 𝗵𝗮𝗿𝗱𝗲𝗿 𝘁𝗼 𝗽𝗿𝗲𝗱𝗶𝗰𝘁: some things land in a week, others can take 3x as long. 𝗦𝗲𝗽𝗮𝗿𝗮𝘁𝗲 𝗲𝘅𝗽𝗹𝗼𝗿𝗮𝘁𝗶𝗼𝗻 𝗳𝗿𝗼𝗺 𝗿𝗲𝗹𝗶𝗮𝗯𝗶𝗹𝗶𝘁𝘆: It’s fine to ship fast and messy behind flags, until you’re touching something mission-critical. Then you need evals, guardrails, and patience. In the end, the conversation made one thing clear: in the AI era, product strategy is less about predicting the next three years and more about making good bets, quickly, with imperfect information—and building teams who are not only capable of that, but genuinely excited by it. Watch the highlights below, and subscribe to the Fin Luma calendar to join future meetups in your city.

Fin is now officially certified under AIUC-1, the new global standard for responsible AI Agents. This independent, enterprise-grade certification validates Fin’s safety, reliability, and resilience, giving every business confidence that their AI Agent is built on trust. Link in thread 👇

“I can do something now in one day that used to take two months before.” AI is forcing companies to reevaluate everything, and at our Fin x @glean x @heyjasperai meetup during @WebSummit in Lisbon, @destraynor, @jainarvind, and @timyoung explored exactly that – from how they build, sell and price products, to how they build thriving teams in the AI era. Here are some takeaways that stood out: 𝗬𝗼𝘂𝗿 𝗼𝗹𝗱 𝗼𝗽𝗲𝗿𝗮𝘁𝗶𝗻𝗴 𝘀𝗽𝗲𝗲𝗱 𝗶𝘀 𝗻𝗼𝘄 𝗮 𝗹𝗶𝗮𝗯𝗶𝗹𝗶𝘁𝘆 As Arvind put it, our “muscle memory has been built to operate at a speed which is no longer the right speed anymore.” Work that took two months now takes a day. Teams that do not reset their pace will feel permanently behind. 𝗙𝗲𝗮𝘁𝘂𝗿𝗲𝘀 𝗮𝗿𝗲 𝗶𝗿𝗿𝗲𝗹𝗲𝘃𝗮𝗻𝘁 𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝗼𝘂𝘁𝗰𝗼𝗺𝗲𝘀 Buyers are tired of shiny demos. They want specific business problems solved, with clear KPIs, not just AI features. Timothy said it plainly, “We’re not selling you code. We’re not even selling you access, right? We’re selling you business outcomes.” 𝗘𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻 𝗮𝗻𝗱 𝘃𝗲𝗹𝗼𝗰𝗶𝘁𝘆 𝗮𝗿𝗲 𝘁𝗵𝗲 𝗿𝗲𝗮𝗹 𝗮𝗱𝘃𝗮𝗻𝘁𝗮𝗴𝗲 In a world where surface features can be replicated quickly, speed, persistence, and raw product momentum are what separate long-term winners from short-lived copycats. Take a look below for the highlights, and subscribe to the Fin Luma calendar to join future meetups in your city.