Intercom

Pioneer 2025 brought together industry leaders and innovators to explore the future of AI and customer service. We announced Fin 3, the most powerful AI Agent for customer service, capable of handling complex queries across every channel. We shared our vision for what comes next: Fin evolving into a single, unified Customer Agent that can handle the entire customer experience. We heard from customers and partners leading the way in AI-first support, shared insights from product leaders, explored what’s next for human and AI collaboration, and capped it all off with a powerful moment of clarity about the scale of the opportunity ahead from none other than Gary Vaynerchuk. Watch the highlights below.

Hallo, Berlin. 👋 We’re opening our next R&D hub right here in Germany’s capital – Berlin. This expansion marks a big step toward realizing our vision: turning Fin into the world’s leading Customer Agent; not just a Customer Service Agent. Our Berlin office will become home to a growing team of engineers, AI researchers, data scientists, designers, and product builders – 100 people over the next year – all focused on advancing Fin and shaping the future of AI for customer experience. Berlin’s rare mix of technical excellence, creativity, and energy is the perfect place to help us push the boundaries of what AI can do for customers. If you’re an ambitious builder ready to work on one of the most exciting AI Agents on the market, this is your moment. Learn more about our Berlin expansion and open roles in the comments ↓

AI reshapes how value is created, measured, and communicated across the business. The “Economics” chapter of The AI Agent Blueprint was developed with insights from leaders like Vijay Mehrotra and Nick Clark, who are helping redefine how businesses should think about metrics, ROI, and value in the age of AI. If you’re a CS or CX leader looking to align with finance, justify reinvestment, or evolve how support contributes to company growth, the Blueprint will help. Read it now. Link in the comments 👇

“ROI isn’t just about cost, it’s about creating capacity.” – Guneet Singh, XMP Recorded live at Pioneer 2025 in New York, this episode of The Ticket podcast with Guneet Singh and Leanne Harte looks at how AI is reshaping what return on investment really means. As Guneet explains, it’s a shift from efficiency to effectiveness, and it’s redefining what great customer service looks like. Watch the full podcast from Pioneer 2025 below. 

"Are you squeezing this opportunity hard enough?" That was GaryVee’s message at Pioneer 2025 to a room full of people already deep in the work of applying AI, while many organizations are still figuring out where to start. Customer support teams, in particular, find themselves in a unique position. Not just adapting to AI, but actively shaping how it gets used. Implementing it. Running with it. The knowledge these teams hold – where AI works, where it fails, where it needs help – is no longer just operational. It’s strategic. It puts support leadership closer to the center of company-wide AI decisions, alongside CTOs and CIOs. Watch the clip. Then consider the question again: Are we squeezing the opportunity hard enough?

As Fin takes on more complex queries and works across more channels, visibility and control become even more important. We've made significant improvements to our Insights product to help you better understand and improve Fin's performance at scale: → CX Score now includes reasons: learn why a conversation went well (or didn’t), from product feedback to answer quality, so you can truly understand the customer experience. → Topics Explorer offers a Topic Trends report that surfaces what’s changing week to week – volume spikes, drops in resolution, emerging issues. Our Topics Explorer already used AI to organize conversations into topics and subtopics, but you now have full curation control to rename, merge, move, and create new ones. → AI-powered Suggestions are now smarter. It learns from when you accept or reject suggestions, spots duplications and contradictions, and recommends updates to data, actions, and guidance. This is possibly the biggest leap forward in helping you answer more and more complex queries. The result: a complete view of how Fin is performing, and concrete insights you can use to make the customer experience as good as possible. See these new Insights updates in action at Pioneer 2025 below.

ChatGPT now connects with Intercom. Starting today, OpenAI's ChatGPT can pull real customer context – conversations, tickets, and user data – directly from Intercom. That means your product, marketing, and sales teams can bring the customer’s voice into every decision. → Product can spot bugs faster. → Marketing can refine messaging. → Sales can prep QBRs with real insight. Learn more about how to connect at the link in the comments ↓

At Pioneer 2025, Paul Adams called Fin live on stage. No scripts, no Hollywood edits. In ~90 seconds, Fin verified identity, pulled up account data, handled an interruption, offered options, resolved the issue, and followed up by email. Behind the scenes, Voice has had major upgrades: better Guidance, customization, testing, and summaries – and it works with all major telephony providers. The more channels you deploy Fin on, the more work it can take off your people, and the better the customer experience becomes. Fin already supports more channels than any other AI Agent, and we’ve just added two more: Slack and Discord. In Slack, Fin fits into the team workflow; threaded replies, emoji reactions, real names and avatars. Discord works the same way, built for the audiences that live there. Watch below to see what happened when Paul (and Fin) picked up the phone.

Most AI Agents can answer basic questions. But it's the complex ones that really move the needle. That’s why we built Procedures. It’s one of the most powerful parts of Fin 3. It lets you train Fin to resolve complex queries (refunds, billing, troubleshooting) from start to finish. Here's how it works: ❶ First, describe what you want Fin to do in natural language. ❷ From there, layer in precision where you need it most with deterministic controls. ❸ Then, Fin approaches every step agently, reasoning about what to do next rather than marching through a fixed path. ❹ Last, we've built an AI Assistant to help you configure all of this. Describe the process you want and it will draft a first version for you. The result: a simpler, smarter way to automate the hardest parts of customer service – without sacrificing the natural experience customers expect. Read more about Procedures in this article from our VP of Customer Support, Declan I.

We've added significant agentic power to Fin with our new Procedures feature, and that power also adds uncertainty. How well will Fin follow a Procedure? How well will it reason? We built Simulations to help you test this at scale. You can run fully simulated customer conversations to see exactly how Fin will behave and reason before deploying changes, and build confidence as it takes on increasingly complex queries. You can define success criteria, run tests across different customer segments, and debug any failures directly in the interface. And there's an AI Assistant built in to suggest changes or recommend additional Simulations to run for more complex Procedures. See how it works below.