Ai Video, UGC is okay, nothing special.
Out of all the AI video generation tools I’ve tried, ElevenLabs is probably the best of the bunch. Unfortunately, that’s not saying much. I’ve spent around £40 on credits and have very little to show for it. Between this week and last week, I’ve wasted approximately 11 hours trying to generate videos that should have been relatively straightforward. The biggest issue is that the platform simply doesn’t follow instructions consistently. For example, I repeatedly asked it to create a vortex effect where pages were sucked into a phone screen. I explained it multiple ways, refined the prompts, simplified the prompts, and even removed unnecessary detail. Instead of pulling the pages into the screen, it repeatedly threw them outwards and dropped them onto the floor. I also tried creating a transition between two images. The task was incredibly simple. Keep everything identical and only transition from one image to the next. Despite numerous attempts and increasingly restrictive prompts, it continued creating new images, inventing new logos, adding random animations, pressing buttons that didn’t exist, and changing things I specifically instructed it not to change. The irony is that the more specific I became, the worse the results often got. The UGC selfie videos were reasonably good. However, even then there were odd moments. In one video, the AI actor randomly quacked at the end for absolutely no reason. I have no idea where that came from. My conclusion is that these tools are fine for generating basic talking head content, voiceovers, and simple visual concepts. If you don’t have a product that requires precision, they may be useful. However, if you’re trying to market an app, create accurate product demonstrations, maintain brand consistency, or produce detailed motion graphics, the experience becomes frustrating very quickly. The technology still struggles with relatively simple instructions and often decides to create its own interpretation instead. After spending money, credits, and hours trying to make it work, I’ve come to the conclusion that it’s often cheaper, faster, and more reliable to create the content yourself. At least when a human edits a video, the result is predictable. At the current price point, I don’t think the value is there. The platform shows a lot of promise, but the amount of wasted credits caused by inaccurate generations makes it difficult to justify continuing to use it. Unfortunately, I won’t be returning as its expensive using something that doesnt work and I am a founder using my own money. My business is not generating any money so Im using my monthly earnings from my main job. This is too expensive for such poor quality