The paradox of creation: When AI toys are trapped in the binary choice between “intellectual property” and “product”, the true winner lies in “creating life”

“Should we start with creating the IP or the product when making AI toys? This is probably the question I have been asked the most recently. Interestingly, almost every questioner, whether they are factory owners with supply chains or creative people with mature IPs, actually fall into a thinking trap the moment they start speaking. Because their questions themselves expose the ingrained logical inertia from the traditional toy era.

In the old industrial landscape, this is indeed a classic Hamlet-like dilemma. Either take the IP-driven approach, like Disney, first create animated films that sweep the world, make Queen Elsa widely known, and then turn the sentiment into princess dresses and magic wands; or take the product-driven path, like Lego, first refine the exquisite building blocks structure, and then give it a soul through comics and animations. These two paths have worked repeatedly in the past decades because they are all based on the same foundation: before the user makes a purchase, they have already marked the price of that ‘already existing thing’ through advertising and reviews.

However, AI toys are irreversibly rewriting this iron law. It is no longer a static doll, but evolving into a ‘continuous relationship’. For traditional toys, the value is locked in the moment of payment; for AI toys, the true value often emerges after each late-night conversation, morning greeting, and long companionship, gradually revealing itself. Therefore, that perplexing ‘should we start with the IP or the product’ pseudo-problem masks a more essential core inquiry: Why do users keep coming back for it?

If this question has no answer, then starting with the IP will merely become the creator’s self-indulgence, even if they draw the most beautiful exterior, it cannot fill the void in the soul; starting with the product will merely become the engineer’s assembly, creating even the most exquisite hardware but unable to establish an emotional connection. Conversely, if this question has a clear answer, the IP and the product will no longer be the starting point, but the natural outcome. The truly successful path that emerges will be only one: start with ‘people’, then with ‘things’.

Looking back over the past year, we have reviewed a large number of failed projects and found that those teams that ‘started with the product’ often died from wishful rationality. They are accustomed to first determining a physical form, whether it is plush or hard, then connecting it to a general large model, setting a seemingly reasonable price, and finally putting it on the shelves for sale. This sounds logically coherent, but it ignores a fatal flaw – they have only created a ‘function assembly’, not a ‘character’. User feedback is often polite ‘It’s quite interesting’, but the subsequent data is cold ‘Not really needed’. The root cause is that the product lacks that most crucial soul: Why should users establish a relationship with it? If it is just a talking box, a voice assistant in the phone can do it too, why go through the trouble?

Of course, upon hearing this, many content-originating practitioners tend to take the other extreme: crazily building a world, writing dozens of pages of character settings, and drawing exquisite promotional posters. However, this often falls into the ‘content trap’. An IP is not drawn, nor is it written, but ‘used out’. Without a specific product as a carrier that allows users to touch, talk to, and interact with it every day, then even the most grandiose setting is just an empty fantasy. An IP detached from the product scene is like a tree without roots, and will never take root in the user’s heart.

So, what is the correct path? In the AI toy industry of 2026, it is verifying a brand-new, backward-looking product definition logic. This is no longer a simple linear process, but a process of sculpting from the soul to the body.

First and foremost, the most crucial step is not to design the IP, but to define ‘people’. You must take the time to answer this question: What kind of existence is this AI toy? Is it a mean but kind-hearted friend who always hits your sore spots at critical moments? Or is it a guardian with extremely stable emotions, who will gently comfort you no matter how you break down? Or is it a fellow who shares the same frequency and accompanies you in your madness, never judging you? This is the true “personality” setting, which is far more important than appearance.

Secondly, after establishing the “person”, don’t rush to think about what it can do, but define the “relationship”. What you need to think about is why users keep seeking it? What is the specific form of this relationship? Is it the sense of ritual of saying goodnight before sleep? Is it the dependence of instinctively wanting to hold onto it when feeling down? Or is it the first person you want to confide in when you want to make a joke? This “relationship entry point” is the core competitiveness of the product, which determines whether users “discard it after use” or “have it for a long time”.

Finally, when the “person” and “relationship” are clearly defined, you can enter the traditional “product design” stage. Is it made of plush fabric or hard materials? What is the size? Does it need a screen? What AI solution is adopted? How much is the price? You will find that the product form at this time is “reversed” by the logic of the previous two steps, rather than being decided by the founder sitting in the office and making a snap decision. This product, which is “reversed”, inherently has the gene for establishing a relationship with users.

Why must this be done this year? Because the industry trend has changed. In 2025, users are still in the “curiosity period”, as long as the product can have a few conversations, it can attract purchases. But in 2026, users have become realistic and picky. Facing the dazzling array of AI toys on the shelves, they will instinctively ask, “Why do I have to keep using you?” This soul-crushing question is enough to kill 80% of the products lacking personality charm in the market.

If you are currently preparing to enter the market, or are stuck in the middle of the project, I suggest you temporarily forget the binary opposition of “IP” and “product”. If you come from a factory or supply chain, please be sure to restrain the urge to directly make the product, and first complete the courses of “personality” and “relationship”. If you come from the content or IP field, please force yourself to implement, find a minimum viable product to verify your setting. A reasonable time allocation is: spend one day thinking thoroughly about “person + relationship”; then spend seven days quickly making a product prototype that can bear this relationship. The sequence must not be reversed.

Finally, I would like to repeat that sharp sentence again: If you are still debating “should I do the IP first or the product first”, it is better to directly ask yourself a more fundamental question: Why do users stay? As long as this question has no answer, doing the IP will be empty and doing the product will lead to death. In this new era, the greatest product manager is no longer the one who is most familiar with mechanics or code logic, but the psychologist who understands human hearts and is best at “creating people”.