Last night at 10 p.m., a client asked for an ad where “the audience chooses the plot”
I stared at a blank document and immediately priced out editing, voiceover, and animation. The biggest fear: doing all that work, only for the client to dislike the result. That kind of job—where they understand the idea but can’t quite picture the final piece—can easily swallow an entire night. I’ve also gotten stuck here: I asked a pile of professional-sounding questions and ended up making the client even more confused.
What is Genie 4? It turns text into an explorable world, not an ordinary text-to-video clip
Basically, you describe a scene, and it generates a virtual world you can move through, interact with, and try actions in. Google has placed it in Project Genie, currently available to Google AI Ultra subscribers. Google’s demos have included interactive worlds such as Land of Leaps: users aren’t watching a fixed video; they walk around, find clues, and change the outcome themselves.
There’s already a real person named Priyanka? Sorry, don’t take that as a real name. I know an independent designer, Lin Yue, who runs a studio in Hangzhou and recently tried a similar route. Last Saturday afternoon, she was developing an event concept for a family restaurant. Instead of delivering a finished ad first, she generated an interactive sample where “a child opens the door and chooses a different task,” then let the client play for ten minutes. Her words: “At least we’re discussing something we actually saw, not an effect I imagined in my head.” Still, an official demo doesn’t mean ordinary accounts can reproduce it reliably. Complex actions and continuous storylines still take repeated attempts.
What would it cost to try it today?
Money: the public entry point is a Google AI Ultra subscription, officially priced at $249.99 per month (roughly RMB 1,800), with regional pricing and taxes possibly changing. Time: I’d start by setting aside 60 minutes, aiming only to validate one scene—not produce a finished piece. Technical barrier: “being able to describe the scene clearly in Chinese” is enough, although people without eligible access may not get in yet. First step: open Google’s Project Genie page, check whether your account has an experience entry, and, if available, enter a simple scene such as “a convenience store on a rainy night; the camera enters, and a letter appears behind the shelves.” If the subscription entry point isn’t showing, take note and don’t rush to buy a plan just to test it.
I’d decide whether to try it based on where my business is now
If I’m just getting started and can’t write complex prompts yet, I’d storyboard with a free design tool first rather than pay for Genie 4 right away. If I have one or two clients and keep hitting requests for interactive content, I’d use a low-cost concept sample to clarify the brief. If I’m scaling and already have a video team, I’d track generation time, revision count, and rights ownership before deciding whether it belongs in the production workflow. Not everyone needs this tool, and waiting is fine. For me, validating with a cheap sample that the client genuinely wants interactivity comes first.