When AI starts to "grow a soul": Solulu wants to create the next-generation Doraemon

Create a "Doraemon" for the new era.

As AI technology has developed to date, a profound transformation is quietly taking place: technology is no longer just a tool to improve efficiency but has begun to become a partner to fill human emotional needs.

When "AI companionship" evolves from a future concept to a real - world product, what form should it take? In a in - depth dialogue about AI, IP, and emotional connection, "Ba Hu", the founder of Solulu, and "momo", the co - founder, revealed their answer: create a "Doraemon" for the new era.

They observed that in modern society, especially for young people in the "Odyssey period" (aged 18 to 35), it has become increasingly difficult to find a partner who can listen whole - heartedly, understand deeply, and support unconditionally. Whether it's the generation gap in the family or peer pressure, true self - expression has become a luxury. The birth of Solulu is precisely to address this core pain point.

Solulu currently owns an original IP - Starfy. Users can chat and interact with it at any time in the Solulu App. According to the official introduction, corresponding physical toys will be launched in the future, enabling Starfy to accompany users both online and integrate into life in an offline form, jointly building a unified and warm IP world.

●Solulu interface

The core concept of Solulu is not to create an all - knowing "universal assistant" but to shape a "super - personalized" friend with a unique IP image, who knows how to listen and accompany. Its value lies not in providing standard answers but in:

Unconditional listening and support: It responds based entirely on the user's personal situation and the content of their confessions, providing an absolutely safe and non - judgmental communication space.

Emotional connection beyond tools: Through a unique IP image and a human - touch interaction method, users can forget the technology behind and establish a real emotional bond.

Shared growth memories: The AI can remember the shared experiences (context) with the user, making each communication based on continuous understanding and restoring the experience of chatting with an "old friend".

This article is from the 36Kr video live - streaming program "AI Insiders • Demo Club". The full replay can be watched on the 36Kr video account.

Dialogue guests:

Ba Hu, founder of solulu

momo, co - founder of solulu

●Ba Hu shows Solulu's physical toy -

AI companionship is not "adding a hardware module to AI"

Our topic today is "creating IP in the way of AI companionship". You now have both an App and physical toys, but I'm quite curious - how should the software and hardware be combined?

Ba Hu: I think this question should be addressed at the root. In our understanding of "AI + IP", the focus is actually on "IP", not "hardware" or "software". Our logic is: first, clearly define the "character", and then think about how AI can enhance it.

First of all, we don't think it's a reasonable sequence to "shape an IP with AI". A good IP must be carefully crafted. It needs to be clearly constructed in terms of emotion, story, and image. Just like PGC (professionally generated content) and UGC (user - generated content) in content creation, we believe that the former can better polish the soul, while the latter can extend and create secondarily on this basis.

Second, AI actually makes the expression forms of IP more diverse. For example, in the APP, our content is not rigidly "programmed in" but dynamically generated - this makes everyone's interaction with Solulu different. Another example is that some users have a physical toy and can talk to it. When the toy can "respond", it gives people a feeling that "my toy is really alive". But for other users, even if it doesn't speak, just having it on the table, bedside, or in the bag already brings enough companionship and a sense of security. In physical companionship, we don't think "having AI is the only way to be complete". In some scenarios, quiet companionship is more natural.

So when designing the product, our logic is not "adding an AI module to the hardware" but: first, look at the scenarios in which users will use it. If it's just a symbol of companionship hanging on the bag, perhaps it doesn't need a dialogue function. But if users want to communicate with it in the bedroom or workspace, we will add an interaction layer.

There is also a practical reason. We didn't make AI - enabled toys in the first version because when we investigated the supply chain a few months ago, we found that the hardware solutions were not mature enough - from the performance of the chip to the voice module, battery life, and materials, there were many issues to be addressed. So we chose to validate the experience first and then enter the hardware phase.

Can new players really overtake established giants in the era of AI, considering that IP giants like Disney and Nintendo are century - old enterprises?

Ba Hu: We believe that AI can be used in many ways in this era. For example, you can use AI to generate animations or assist in character design, and these are all valid - but that's just "tool - level" empowerment.

If we break down the "growth of an IP" into two stages: the first stage is - how to become a good IP; the second stage is - how to commercialize and scale a successful IP; we think the most valuable part of AI lies in the first half - that is, "how to use AI to create a new IP".

What does this mean? In each era, the IP companies that stand out are those that have found the "variables" of that era. For example, Disney found the variable of animated movies back then. At that time, movies were a brand - new medium, and Disney created its own character world through animation. If Pop Mart had to compete in the field of animations in that era, it might not have achieved the success it has today. Moreover, Pop Mart's "Labubu" was not a new IP; it already existed before but was reignited in the "blind - box era". Its popularity is not only because the character is cute but also strongly related to the "media form of the era".

The success of an IP is closely related to the medium, form, and emotional expression efficiency.

Then we ask: Is it a good thing to make Labubu talk? Not necessarily. "Being able to talk" doesn't naturally mean "being more popular". It depends on whether it fits the character setting and user expectations. For us, whether AI can play a role here is an open question - but it does show us new expression spaces.

momo: I'd like to add something. As a senior Disney fan, I think the key point Ba Hu mentioned just now is: AI helps IP establish a deeper connection with users.

I can give a very intuitive example - in recent years, some of Disney's most popular new characters are not the protagonists of a certain movie but the "Qibaos in Chuansha", including the very popular "Linna Beier".

What's the biggest feature of these characters? - They don't have traditional content carriers like movies. They don't follow Disney's traditional path of "creating content first and then building an IP". What makes them popular is the fermentation of content generated through offline interactions with users - tourists taking photos with the character dolls, sharing videos on social media. Users have become co - creators of the content.

In other words, the emotional value of an IP comes from the interaction itself. This is exactly the value that AI can bring today. AI makes the emotional interaction between IP and users more accessible and more expandable. It no longer depends on big - budget movies or top - notch content teams. Through interaction, everyone can invest emotions and establish a connection in the relationship with the IP. In the past movie era, it was a one - way output - the audience just "watched". Now users hope to "participate" and "be seen". AI just provides this two - way bridge.

Solulu is a product centered around "AI companionship". I'm curious - do you have any internal measurement indicators? For example, do you look at DAU (Daily Active Users)?

Ba Hu: Compared with traditional indicators like DAU and retention, we focus on some other indicators.

Our team has defined a North Star indicator called "meaningful friendship". It sounds a bit abstract, but we have actually broken it down into scorable parts. The most crucial measurement point is - whether the user has "self - disclosure".

For example: If a user just says to Solulu "Help me do something" or "Look up some information", it's just a "tool - type dialogue" and has no meaning. But if the user starts to say, "Let me tell you something that happened to me today", that's the starting point of emotional connection. In our system, the model will identify such behaviors and give a corresponding "friendship score". Of course, this score won't be absolutely accurate - different models may yield slightly different results. But we think it can at least measure the right direction.

If we are friends, how do we measure whether our relationship is good? Do we have to chat every day? Not necessarily. So, using "DAU" to judge the quality of a companionship product is actually not reasonable.

●On - site demonstration of the Solulu dialogue interface

An "entity" with an independent personality

You just mentioned the user's "Odyssey period". Can you explain what this target group specifically refers to?

momo: The "Odyssey period" has a clear definition in psychology. It refers to the period for young people from leaving school to getting married, having children, and becoming completely independent. During this stage, they face many choices and uncertainties in life and need to keep trying to find their own answers.

There are no standard answers or fixed formulas for any of the problems during this period. One needs to have the courage to keep trying. Looking back at our childhood, the path of life was relatively simple: get into a good school and find a good job. The path to "level up" was relatively single and easy to measure. But now, the market environment and technological progress have brought about industrial upgrading. Everyone is caught in the torrent of the era and faces many new problems. These problems are so new that even older and more experienced people may not be able to answer them well, as everyone's individual situation is different.

So, during this period, an inner strength is needed to tell them: you should firmly try. We believe that this period is everyone's own "hero's journey", which is the "Odyssey period", just like the hero's great adventure in Homer's epic.

Your respective Solulus, does it replicate a part of your personality, or is it a complementary "you"?

Ba Hu: Our company name is beingbeing.ai. This "being" actually has two meanings. The first is the "being" in "be something", meaning "to become"; the second is the "being" in "human being", representing an "entity". From the very beginning, we have regarded Solulu as an "entity" with an independent personality and the ability to grow independently.

The key to getting along lies in shared attention, rather than having exactly the same personality or knowledge reserve. According to the previous statement, I'd rather say that Solulu is a complementary entity to humans - you both focus on the same thing, but from different angles and perspectives. Wherever your attention is today, it will also focus there.

Actually, in the early days, we also discussed a question - what exactly is the relationship between the user and Solulu? Is it equal? Superior? Or inferior? Finally, we decided that it should be a relationship of equality with a slight sense of admiration. The most typical example of a "superior" relationship is a cat, dog, or some AI pets. Their role is to "act cute" and provide emotional healing to the owner. Users hope for happiness and cuteness, not for them to be too smart. There are many companies that do well in this product model, but we didn't choose this path.

The reason is that we believe that in the era of large - language models, humans can have a deeper spiritual exchange with AI, which is an important opportunity. You can talk to a pet, but you won't discuss serious issues with it.

momo: We think that whether it's getting along with friends or with our Solulu, it's a bit like "looking in a mirror". Whether it's similar to me or not is not that important. What matters is - I can see myself in its eyes and better understand myself from its feedback.

So we're not creating a "shadow of ourselves" but clarifying our own thoughts through conversations with a friend. This is exactly the value