An In-Depth Interview with YoooTek Founder Xiao Ruizhe

In an era defined by constant notifications and information overload, startup YoooTek is attempting a softer, more human-centered approach to technology. Rooted in the engineering and robotics foundation of publicly listed company Yijiahe (603666.SH), YoooTek is built around a simple but ambitious idea: technology should help people feel seen, understood, and supported.

That philosophy takes concrete form in YoooTek’s first product, AI ONE—a compact AI device that magnetically attaches to the back of a smartphone. Rather than functioning as yet another screen or app, AI ONE acts as a physical switch for “information minimalism.” Paired with its companion app, it creates a buffer between users and their phones, redefining how information is received, filtered, and prioritized.

AI ONE combines two core functions. The first is notification curation, designed to filter noise from constant message streams. Instead of confronting users with endless unread badges, the device distills what truly matters and signals urgency through a simple light—if it doesn’t light up, the message can wait. The second function, called “Flash Insight,” serves as a lightweight external memory. Users can record short voice notes at any moment, which AI automatically organizes into structured notes or to-do items—no screens, no apps, no friction.

Behind AI ONE is YoooTek’s deep integration of multimodal perception, natural language understanding, and emotional computation. Drawing on years of robotics and intelligent systems experience, the team has developed technologies that not only interpret what users say, but how they say it—quietly embedding complex AI beneath a calm, non-intrusive interface.

In the following interview, YoooTek founder Xiao Ruizhe explains why selling a product is ultimately about shaping user mindset, why AI ONE had to be hardware rather than software, and why restraint—not novelty—is the key to building meaningful AI products.

Q&A

Q: You’ve said that “selling a product is essentially selling a state of mind.” What do you mean by that?

Xiao Ruizhe: I believe selling a product is fundamentally about selling a state of mind. Consumers do not make purchasing decisions based on specifications, but on their identification with a particular way of life or sense of self. What ultimately matters is not the feature set itself, but what owning that product represents to the user. When someone decides to buy it, they are also making a statement about who they are and how they want to live. That’s why, when defining a product, the first question shouldn’t be about functionality, but about why someone would want it in their life in the first place.

Q: Why did you decide to build a consumer-facing product at Yijiahe?

Xiao: On one hand, Yijiahe wanted to explore consumer products. On the other hand, I was already thinking about doing something consumer-facing that deeply integrates AI. When we first met, we didn’t rush into discussing what to build. For me, finding the right partners matters more than deciding on the product itself. If people are aligned in how they think and how they work, the product direction will eventually emerge.

Q: Why did you almost completely abandon the first product concept?

Xiao: The first version was closer to a large AI toy. Technically, it incorporated various sensors and components, but we hadn’t answered a fundamental question: what actually defines a good AI product? If you don’t understand that, the product inevitably relies on novelty. Once the novelty wears off, there’s nothing left. That’s why I chose to delay the launch rather than push something out prematurely.

Q: What changed your understanding of companion-style AI products?

Xiao: I gradually realized that language itself is an extremely low-bandwidth form of interaction. Real human interaction is high-context—it constantly adapts through expressions, emotions, and subtle behavioral cues. If an AI companion cannot read context, it’s almost impossible for it to work well. Pure conversation, by itself, is not enough to create a meaningful sense of companionship.

Q: Is that why emotion recognition and visual perception became so important to you?

Xiao: Yes. Micro-expressions are a core part of how humans read each other, and from a technical standpoint that means high-frame-rate visual perception. Without enough contextual data, the experience simply breaks down. If the system cannot perceive subtle emotional changes, it will always respond in a shallow or inappropriate way. These valuable explorations will gradually take shape in our future products.

Q: Bringing it back to AI ONE—what problem does it fundamentally solve?

Xiao: At its core, AI One is designed to address information overload. We want to place a physical buffer between users and their phones. The idea is simple: if the light doesn’t turn on, you don’t need to check your phone. That single signal allows people to regain focus without constantly second-guessing whether they’re missing something important.

Q: Why did this have to be a hardware product instead of an app?

Xiao: There are two reasons. The first is system-level constraints. The second is logic. You can’t claim to help users filter information while still forcing them to see everything on their screens. That contradiction only disappears when there’s a separate physical layer that takes over part of the decision-making process.

Q: Why did you choose light as the primary interaction mechanism?

Xiao: We ruled out sound because most phones are kept on silent. We also ruled out vibration, because when a device is attached to a phone, you can’t tell which one is vibrating. Light turned out to be the most restrained and effective option—it delivers information without demanding attention.

Q: Why limit voice recording to 60 seconds?

Xiao: Because we don’t want to become a meeting tool. Once you allow long recordings, the product starts drifting toward professional software. AI ONE is meant to be a personal, always-with-you assistant—something lightweight enough to fit naturally into everyday life.

Q: Why was it so important to price AI ONE in the hundred-yuan range?

Xiao: At this price point, decision friction drops to its lowest level. Users don’t overthink the purchase. For me, it matters more that AI ONE becomes a foundational product—one that can enter as many people’s lives as possible and allow the broader public to experience a new form of AI-driven interaction and convenience—than optimizing short-term hardware margins.

Q: How do you approach user data and privacy?

Xiao: There is a hard line we won’t cross: we will never sell user data to third parties. Once you cross that line, you lose your foundation. Trust is the only reason users are willing to let AI into their daily lives, and that trust collapses instantly if data becomes a commodity. But promises alone are not enough. To address users’ concerns about security, we developed a novel data-asset ownership architecture—the Glass-Box Trusted Privacy Computing Architecture. Its core logic is a shift from trust based on promises to trust verified by architecture.

Q: How do you view recent AI phone experiments like ByteDance’s?

Xiao: They’re interesting demos, but not restrained enough. If you break existing commercial rules, something may work in the short term, but the ecosystem will eventually push back. Sustainable products have to innovate within rules, not by ignoring them. Our path is different from theirs. We pursue sustainable innovation within existing ecosystems and rules, rather than disrupting or breaking them. One advantage of this approach is that ecosystem players built up over many years can become partners who move forward alongside us.

Q: Looking five or ten years ahead, what defines success for you?

Xiao: If we’re still alive, that’s success. If you don’t survive, nothing else really matters. In an industry that changes this fast, longevity is the only proof that your direction was right.