I tried Luma AI with high expectations, especially since I wanted to participate in their contest and produce a commercial to present at the Cannes Festival. Unfortunately, the experience was extremely disappointing.
To be honest, the platform is simply unable to generate what you ask for. Even with detailed prompts and multiple attempts, the results are inconsistent and far below the level expected from modern AI video generation tools.
Luma is years behind tools like Veo 3, Kling, and Seedance. Those models can produce coherent scenes with minimal prompting — sometimes just from a single reference image — while Luma struggles with the basics.
The issues are constant:
* Characters are not consistent from shot to shot
* The cinematic style constantly changes
* Lighting and atmosphere shift randomly
* The overall artistic direction is unpredictable
Even when trying to heavily guide the model, it cannot maintain visual consistency across scenes. For a platform that is supposed to help create video content, this is a major flaw.
The interface doesn’t help either. The platform is not intuitive and feels immature and poorly designed. Instead of focusing on creating, you spend most of your time trying to fix errors or regenerate unusable outputs.
Honestly, with this level of quality, it’s hard to imagine using Luma for any serious project — let alone for a commercial meant for something like the Cannes Festival.
At this point, the platform simply offers no real value compared to far more advanced competitors.
Claim your business profile now and gain access to all features and respond to customer reviews.
