Chen Zhang’s favorite project was helping Paramount develop animation for Bumblebee, one of the characters from the Hasbro collector toy franchise Transformers. Her startup Aquifer went to work to make the yellow Autobot dance. As the co-founder and CEO of the AI-driven animation company, Zhang likes to dispel the notion that artificial intelligence removes the effort to create content — or dancing robots. She notes that AI can give you a jumpstart — but creators are behind every successful product.
You can hear more from Zhang in the video below or read our lightly edited transcript.
TRANSCRIPT
Chen Zhang
Aquifer is an all-in-one instant animation platform designed specifically for the needs of brands and IP.
We are a software technology platform powered by AI that allows brands to create very quickly animated videos featuring their mascots, their characters for social media, or episodic content marketing materials, in a matter of minutes, instead of days or months.
The way that it works is every brand on inside of Aquifer has gone through a brand onboarding process in which their characters are ingested into Aquifer, and we develop what is called a custom style profile, meaning it is a profile that informs all of the movement data that is generated inside of Aquifer. An example that I often like to give is the way that a cartoon lizard, says hello would look very different than the way a cartoon version of a little boy or girl might say that. After that process, the brand or whoever’s using Aquifer can then jump in, upload audio, or record audio, and from that audio, they’re able to generate a first pass of face animations, body animations lip sync, and then from there, they can use a very simple interface to manipulate that movement.
When you have a team of animators, they usually start with storyboards of how this character is going to move throughout this specific scene. Then from those storyboards, there’s a script, the lines get recorded, and then animators are using a 3d animation platform like Maya, or they’re doing in a real time platform like Unity or Unreal, and they’re literally creating technical animation data through mocap [motion capture], and manipulating movements, often frame by frame. But it’s a highly technical process that requires skilled people to go in there and generate that, and then from there, they’re exporting animation data, and then eventually, it’ll get rendered out into a video.
What Aquifer is doing is, it’s compressing that entire process, and taking away the most technical components. It can drop animation time, from hours, days, weeks, down to seconds and minutes. It dramatically reduces costs, it dramatically reduces the number of people that have to work on a particular project, to call it done, and so all of that reduces cost significantly.
Some of the fear that people have for AI taking over or taking people’s jobs. I think the best analogy is it, we’ve had machines that have automated stuff for us that used to be very arduous and are now quite fast, and it’s made things more efficient. I think ChatGPT gives the impression that that it is replicating the work that people could do. But I think oftentimes, what’s happening is it’s giving a result, that in order for that result to be usable, there still has to be quite a bit of input from the user. Everyone has been, ‘Oh [ChatGPT] has generated some copy for me.” Well, it generated some copy. Whether or not you want to include that coffee in any front-facing material is another question. I fully believe that it will become a tool like any of the other AI tools we’ve been using.
My favorite brands that we’ve ever had the opportunity to work on, would be Transformers’ Bumblebee. It is a character I’ve loved for a very long time. He emotes through random music or other sound effects, and so it was such an exciting opportunity to be able to work with the Paramount team to generate these Bumblebee animations that they eventually put on their TikTok. He just looked so fantastic. The way he came out of Aquifer, we captured all of that shiny robotic energy and he looks so good.
There’s so many pieces of advice that I’ve learned from lots of wonderful people and just from my experiences in general along the way. I think first and foremost make sure it’s something you love, because it will be something you are thinking about 24/7 and working on. We started this company in 2020, and so we actually graduated from our first accelerator in March of 2020, right as the pandemic was happening, and it was a time when no one knew what was going to happen. If you don’t truly love the subject matter that you’re in, and truly love the day-to-day of working on this problem and working on the product, big shifts like that will absolutely be difficult to go through. Are you still delivering value? How do you deliver that value? And do you love what you’re doing? That can take you a really long way.
I really love solving problems. There is a high that I get from solving a problem that just feels incredible. I don’t know if that can be small or big. It can be as straightforward as we’ve been working on a particular feature, and all the approaches we’ve taken haven’t made sense. Then just one day, for some reason, in a group conversation, we unlock something an insight an idea, a mechanic and it just works. I feelthere’s no better feeling in the world, it just it feels incredible. That’s what I love most about my job. It’s looking at problems and as an entrepreneur, it’s just problems everywhere. There’s macro problems and micro problems. And then being able to use every bit of creativity and communication, and every resource you have to try to solve it and when you do solve it, it is so satisfying.
