John Kim, Co-founder and CEO of Sendbird – Interview Series

John Kim, Co-founder and CEO of Sendbird, is a seasoned entrepreneur who has spent more than a decade shaping how businesses communicate at scale. Under his leadership, Sendbird has grown into the leading omnichannel AI agent platform for enterprises, trusted by brands like DoorDash, Match Group, and Paytm. Backed by ICONIQ, SoftBank, and Tiger Global, Sendbird combines AI automation with human oversight to power more than 7 billion conversations each month. Before founding Sendbird, Kim launched and sold Paprika Lab, a successful social gaming company acquired by GREE, and was once Korea’s top-ranked professional gamer in Unreal Tournament. He also serves as General Partner at Valon Capital, investing in AI, deep tech, and fintech ventures.

Sendbird helps enterprises automate and personalize customer interactions across chat, voice, and video. Its platform blends AI agents with seamless omnichannel communication, enhancing engagement while preserving human context for complex issues. With its reliable, secure, and scalable infrastructure, Sendbird is redefining how companies build lasting relationships through intelligent, AI-driven conversations.

You began your career as Korea’s #1 professional gamer before founding Sendbird. How has that early experience—especially the competitive strategy and discipline from pro gaming—influenced your leadership and vision for the company?

Being a pro gamer probably shaped me more than anything else I did before starting Sendbird. Gaming at that level, every small mistake or improvement makes all the difference in whether you win or lose. It forced me to be disciplined,  practice endlessly and think strategically about what might come next.

One big takeaway was learning how and when to pivot. As a gamer, I hit a point where I realized I was at my peak and that continuing down the same path wasn’t going to take me further. That experience showed me how to spot both limits and opportunities, which ultimately led us to pivot, and that’s how Sendbird was born.

To be transparent, the pressure of competing in front of thousands of people also set me up to handle the pressure of running a startup. When you’ve been in that kind of environment, you start to learn composure under pressure. That mindset – staying calm and tenacious – helped us get through the tough times. In many ways, gaming taught me what I still believe today: persistence, iteration and having the courage to make big pivots will lead to excellence.

Sendbird originally started as a community app for moms that struggled until you recognized an opportunity in in-app messaging. What guided that shift, and in hindsight, what role did listening to user behavior play in transforming Sendbird’s direction?

We started with a simple belief: conversations are the core of every great experience. But our journey to that belief began somewhere unexpected – with a small community app for moms. The idea was to help mothers connect with others nearby who had children of similar ages. The app quickly grew to a quarter-million users, but what stood out wasn’t how they used it; it was why.

Instead of outsourcing a messaging function, we challenged ourselves to build it. Through a hackathon, we created a chat SDK, spun it out and began selling it on the side. Unlike our original app, which had struggled to generate revenue for over two years, the SDK gained traction quickly and brought in paying customers. That moment was pivotal. It showed us the real demand for in-app messaging and validated that listening to user behavior could unlock a much larger opportunity.

In hindsight, the lesson was clear: our users pointed us toward the real value we could provide. By paying attention and acting on those signals, Sendbird transformed from a niche community app into the communications platform it is now. Today, Sendbird powers billions of conversations every month. We’ve helped businesses turn messaging into connection, engagement, and growth. Because conversations aren’t just the start – it’s how they’re delivered, remembered and felt that makes them meaningful.

Sendbird’s AI agents go beyond scripted chatbots: they actively understand intent, hold context, act autonomously, adapt over time, and seamlessly escalate to human agents when needed. How did you design this evolution in agent intelligence, and what were the key challenges in pulling it off at scale?

Our goal in building AI agents wasn’t just to create smarter chatbots, but to enable conversations that feel genuinely human. That required designing a system capable of learning, adapting over time and knowing when to take action on the user’s behalf.

The evolution came from observing real customer interactions. Customers don’t want rigid scripts, constant repetition, or feeling like they’re talking to a robot. We combined natural language understanding with contextual memory and workflow integration to move conversations forward naturally.

The hardest aspect of doing this at scale is balancing intelligence and reliability. It’s not enough for an agent to understand; it also needs to respond consistently, securely and in a way that a company can trust across millions of conversations. We spent a lot of time building the guardrails, training the system on domain-specific data and making sure escalation to humans happens at exactly the right moment. In the end, our approach works on the same principle that guided Sendbird from the early days – listening deeply to how people communicate and designing technology that makes those interactions faster, easier and more natural.

Sendbird enables businesses to engage with customers across in-app chat, SMS, email, social messaging, voice, and video—all through one platform. What strategies did you employ to unify such diverse channels while maintaining both reliability and a consistent user experience?

Unifying channels was never about adding more, it was about removing friction. From the beginning, we built Sendbird with a deep respect for reliability and real-time performance. When you power billions of messages every month, you learn that every millisecond matters and every failure erodes trust.

So when we expanded beyond chat into SMS, email, social messaging, voice, and video, our strategy wasn’t to stitch channels together – it was to architect one conversation fabric. That meant designing a common infrastructure and data layer where every message – no matter the channel – shares the same backbone of delivery guarantees, compliance, and user identity.

On top of that, we built orchestration layers to ensure true continuity. So, a customer can start a chat in-app, pick it up over SMS, and seamlessly move to voice, without ever noticing a handoff.

Our approach was to build from the inside out – one unified foundation instead of a patchwork of channels. That’s what enables us to deliver reliability, consistency, and effortless experiences at scale.

With robust APIs, SDKs, and developer-focused tools, Sendbird doubles down on developer-first design. What made prioritizing developer experience so pivotal—and how has it shaped adoption across industries?

Prioritizing developer experience was essential because our success depends on how easily developers can integrate Sendbird. We’ve built robust APIs, SDKs and developer tools designed to reduce complexity and allow customization at any layer. That approach has made Sendbird a preferred communication platform for teams across industries such as retail, travel, hospitality, and fintech that want to focus on the core customer experience while trusting us for reliable, scalable messaging and AI capabilities.

With Sendbird Desk, enterprises get customizable support dashboards, proactive routing, and combined AI-human workflows. What insights guided the creation of these capabilities, especially around designing AI that boosts both agent productivity and CSAT?

When we built Sendbird Desk, our guiding insight was simple: great customer experiences start with empowered agents. AI should never replace them; it should amplify them.

We saw that most support tools treated AI as an add-on: a chatbot that deflects tickets. But in reality, agents spend an enormous amount of time switching systems, repeating manual tasks, or handling requests the AI could easily prepare or resolve. So, we designed Desk to make AI and humans work side by side.

We designed a real-time AI Command Center – a live workspace where teams can monitor every AI-human interaction, observe performance, and take control instantly when needed. It transforms support agents from ticket solvers into orchestrators and managers of AI agents.

Sendbird is innovating with multimodal AI, expanding knowledge base integrations, and offering versioning tools like prompt history. How are you thinking about the future of AI agents that can process images, files, and structured enterprise content?

We believe the future of AI agents isn’t just multimodal, it’s also multi-presence. The next frontier isn’t only about understanding more formats like images, files, or structured content – it’s about communicating across multiple modes simultaneously. Can you talk and use text chat at the same time? Switch from voice to text without losing context? That’s the kind of seamlessness we’re designing for.

With multimodal AI, agents can see, read, and reason across text, visuals, and data. And with our infrastructure, they can also move fluidly between channels – from in-app chat to SMS to voice – without ever dropping the thread. Imagine being on a support call with an AI agent while texting a photo of your receipt at the same time. The AI instantly reads it, updates your case, and continues the conversation. That’s where Sendbird is heading.

Today, Sendbird empowers billions of users with over 7 billion conversations per month. What were the most significant technical or organizational challenges in scaling from startup to multi-billion conversation throughput across regions and industries?

Scaling to billions of conversations across the globe meant solving two distinct challenges: ensuring every customer interaction is fast and reliable, and maintaining team agility.

We set out to design a system that could handle massive spikes in usage while also meeting strict compliance and performance standards. Equally important was building an engineering culture that remained focused on reliability, observability and constant improvement. That foundation is what enabled the Sendbird platform to scale smoothly across several industries and regions worldwide.

You’ve emphasized personality testing for new collaborators to understand how they work and who they are. How does this cultural lens—combined with your sense of curiosity—shape team dynamics, product creativity, and innovation at Sendbird?

I love personality tests! For me, personality tests are not just about labeling people, but understanding them. I’ve always been deeply curious about how people think and how they approach problems. I ask everyone I work with to take a personality test because it provides a starting point for understanding how they prefer to communicate, make decisions, or handle challenges. At Sendbird, that awareness has been key in how we have built teams and encouraged innovation. Curiosity about people leads to curiosity in our work, and that’s where the best ideas come from.

For example, knowing who’s an INTJ strategist versus an ENFP creative helps us design better collaborations. We can pair analytical thinkers with big-idea explorers. It creates balance, empathy, and mutual respect.

We also believe every AI agent should have its own personality blueprint. Just like people, different situations call for different personalities. The AI agent that helps you file an insurance claim shouldn’t sound like the one that helps you book a vacation.

In your view, what’s next for AI-driven communication platforms? Is the future more about deeper personalization, proactive engagement, agent-to-agent orchestration, or entirely new interaction paradigms—and how is Sendbird positioning itself to lead in these areas?

The future of AI-driven communication isn’t just reactive, it’s proactive. The next generation of customer communication will go beyond waiting for users to initiate a conversation. Instead, AI agents will anticipate requests and proactively surface relevant information to take meaningful action before the customer even asks. That shift is what will truly define intelligent AI-powered communication.

At Sendbird, we’re building the foundation for that future. As we look to 2026, we’re extending our services to create proactive, context-aware agents that can collaborate with humans in real time. We see this kind of engagement as the next major evolution in how companies connect with their customers.

Thank you for the grat interview, readers who wish to learn more should visit Sendbird.