Over the past year, generative AI has shaken the creative industry. Everywhere you look online, you’ll find “AI-generated videos in 5 minutes” or ultra-realistic portraits produced from a single sentence. But what’s really behind the hype? And what can these tools actually deliver in professional video production today? At Createen, a Paris-based creative and video production agency with over 17 years of experience, we’ve always believed that technology should serve creativity — not replace it. That’s why we started experimenting with AI tools as soon as they became available, to explore their real potential and understand their current limits.
What is generative AI?
Generative AI refers to a class of intelligent systems that can produce visual or audio content from text instructions, known as prompts. These models — such as ChatGPT, Midjourney, Runway, Kling, or Veo3 — interpret your text and generate an image, animation, or video that matches your description.
What exactly is a prompt?
A prompt is a textual command that acts as a creative brief for the AI. It describes the scene, camera angles, lighting, emotions, and artistic style. In short, the prompt is the director’s brief: the more precise it is, the closer the result will be to your vision. Advanced users even write prompts in Python syntax, which allows for extremely detailed and structured instructions. (We’ll publish a dedicated article on that soon.).
Text-to-video vs image-to-image vs image-to-video
The biggest beginner mistake is jumping straight into text-to-video — it’s still too random. Instead, start by creating strong static visuals before moving into motion. You have two main options:
We recommend image-to-image, since basing your generation on a real visual ensures a more realistic, brand-compliant result. Once your visual is ready, you can move on to image-to-video, adding a specific motion prompt to bring your image to life.
The best tools right now
Current challenges
The hardest part remains consistency — maintaining the same character, lighting, and background across multiple shots. And what happens when a client wants a change like “make her hair red” or “turn the man’s sweater blue”? Most of the time, you have to regenerate everything. Text and logo tracking also remain unreliable: packshot animations or typography fidelity are still problematic, making full commercial video production with AI alone unfeasible — for now.
The bottom line
Generative AI is an incredible tool — but not a magic wand. Professional-grade results require expert prompt design, multiple render iterations, and strong artistic direction. So yes, experiment, test, and learn. But if you need consistent, high-quality visuals that respect your brand identity, you’ll save time and money by working with a specialized agency. At Createen Agency, we already partner with leading cosmetic and biotech brands to craft AI-enhanced video campaigns that combine creative storytelling with advanced generative technology. We believe the future of visual production is hybrid — a collaboration between human creativity and artificial intelligence.
