When Intercom announced its new R&D hub in Berlin, it sent a clear signal: the city has become a serious destination for companies building at the frontier of AI. With plans to hire 100 people in Berlin across engineering, AI, data science, product, and design, Intercom is betting on Berlin as a key accelerator for the next chapter of its growth – and for the continued scaling of Fin, the #1 AI-powered Customer Agent. 

At the center of that product vision is Thom Rimmer. As VP of Product Design at Intercom, he leads the design of Fin – a product used by over 7000 businesses globally that is on track to pass $100 million in annual recurring revenue. In this interview, Thom talks about what brought Intercom to Berlin, what AI-first design actually means in practice, and how he thinks about building high-performing teams in one of the most fast-moving moments in the history of the technology industry.

Intercom has chosen Berlin for its next R&D hub, planning to hire 100 people across AI, engineering, product, and design. What made Berlin the right city for this step – and what does this move mean strategically for Intercom's growth in Europe? 

Berlin felt like the natural next chapter for Intercom. Europe has always been an important part of our story – both in terms of customers and talent – and Berlin especially has a rare combination of deep technical expertise, a strong AI research community, and a creative, design-forward culture. For a company like ours, where we’re building AI-native products that need both technical excellence and human-centered thinking, that combination really matters. Berlin gives us access to a diverse, international talent pool and a thriving startup scene that’s pushing boundaries in AI – it’s an R&D hub in the truest sense and it will be critical in continuing to scale Fin. We’ll be building some of our core product and AI capabilities here.  

You lead the design of Fin, Intercom's AI Agent for customer service. For someone unfamiliar with the product: what makes Fin different from other AI-powered customer service tools? Where do you see its biggest potential going forward? 

Fin is an AI agent built specifically for customer service – and that focus is important. A lot of AI tools in support are either copilots that assist human agents or chatbots built on rigid decision trees. Fin is different because it’s designed to resolve customer queries end-to-end using a company’s own knowledge instantly, accurately, and at scale. It’s also the highest performing AI agent for customer service and is powered by our own AI models. Fin delivers higher quality answers and resolves more complex queries than any other AI agent on the market, and also works with any helpdesk across every channel, and consistently outperforms every competitor.  

The biggest potential going forward is without a doubt a unified, personalized Customer Agent – an agent that can manage every stage of the customer lifecycle. We believe this is the future of customer service. The Customer Agent will oversee all customer operations beyond customer service and will extend Fin to sales, marketing, eCommerce, and customer success – turning the customer experience into an even bigger strategic advantage than it already is. And with Fin, we are uniquely positioned to be a pioneer in this category.

You've spoken publicly about AI-first design and recently presented at PromptUX in Berlin. How does designing for an AI product like Fin differ from traditional product design – what new skills and mindsets does it demand from a design team? 

Designing AI products like Fin shifts the role of design from shaping static interfaces to shaping systems and behavior. Traditional software is deterministic, you design the flow and the product behaves predictably. AI is probabilistic. The experience depends on how the system interprets context, data, and prompts, so designers are thinking much more about how the product reasons, communicates confidence, and handles ambiguity.  

That changes the skillset as well. Prompting becomes a core design tool, prototyping becomes faster and more experimental, and designers need to be comfortable building working experiences, not just mockups. At Intercom we expect designers to ship code and prototype directly with AI models because that’s the fastest way to understand the interaction.

The biggest mindset shift is around judgment. When machines can generate decent UI instantly, the real value of design becomes taste, knowing what’s worth building, shaping the system’s behavior, and holding a high bar for the final experience. 

Intercom has a strong engineering and AI research culture – your team publishes openly, builds in public, and ships fast. How do you ensure that design not only keeps pace with that kind of high-velocity AI development, but actively shapes it? 

Speed is essential, but the way design influences AI development is by getting involved early and staying hands-on. A big part of our approach is using prototypes and experiments to shape product direction. Instead of waiting for ideas to mature through documents or roadmaps, designers build working explorations that help teams understand what an AI capability might look like in a real product.

We also keep the distance between design and engineering very small. Designers work closely with engineers and often ship code themselves, which creates faster feedback loops and more shared ownership of the product. At Intercom, design isn’t a service function, it’s an equal partner in shaping the roadmap. Our job is to bring taste, clarity, and product thinking into a space that can otherwise become very technology-driven.

Berlin has a dense ecosystem of AI startups, research labs, and creative talent. From your perspective as a design leader: what does Berlin's tech scene offer that other European hubs don't – and what are you hoping to find here?

Berlin has a certain openness and experimentation culture that’s hard to replicate. It’s international, interdisciplinary, and still slightly unconventional compared to more corporate tech hubs. You see strong academic AI research alongside ambitious startups, and a deep creative community that challenges more rigid technical thinking.  

From a design perspective, that is ideal. AI products need rigorous engineering, but they also need bold imagination. Berlin brings both of these together. What I’m hoping to find here are designers and engineers who are excited about shaping the future of AI – not just applying it. People who are comfortable with ambiguity, who want to build foundational systems, and who care deeply about building industry-transforming AI.

Leading design at an AI-native company like Intercom is a different challenge than it was even five years ago. What does it take to build and lead high-performing teams in this environment – and what lessons from your career feel most relevant right now?

The biggest shift is that expectations for designers are much higher now. AI tools can generate decent work quickly, so the value of a great designer is less about production and more about judgment, creativity, and product thinking. Building a strong team in this environment means hiring people who are curious, hands-on, and biased toward action. The designers who thrive are constantly experimenting with new tools, prototyping ideas quickly, and learning in public.

From a leadership perspective, it also means staying close to the work. In fast-moving environments like AI, you can’t rely on heavy process or long planning cycles. Leaders need to help set the bar for quality, challenge ideas, and create the conditions for teams to move quickly while maintaining strong craft. 

Fin is already used by over 7000 businesses globally – including German companies like Ostrom, tado°, and Flink. What does it mean to you personally to build a product that operates at this scale, and where do you see the future of AI in customer service heading?

Building a product at that scale is both exciting and a big responsibility. When thousands of companies rely on Fin to communicate with their customers, the quality of the experience really matters. Small design decisions can affect millions of interactions.

What excites me most is that we’re still at the very beginning of what AI can do. Today AI agents are already resolving a large share of support conversations, but the next phase is systems that can reason more deeply, take actions across more tools, and work alongside human teams much more seamlessly. Our belief at Intercom is that AI agents will handle the majority of customer interactions. The opportunity for design is making those systems feel reliable, understandable, and genuinely helpful, so businesses and their customers can trust them.

Thanks for the great conversation.

Vita: Thom Rimmer is VP of Product Design at Intercom, where he leads the design of Fin, the company's market-leading AI Agent for customer service. Over the past two decades, he has been both a builder and a design leader – scaling teams, taking products from 0 to 1, and embedding design into the strategic foundation of companies across SaaS and agency environments.

Learn more about Intercom’s Berlin R&D office and explore their open roles on their career page.