A creative director asked me for best practices to prompt images/videos with Gemini — what should I tell him?

A creative director I work with asked me: “What are the best practices for prompting images and videos with Google Gemini?”
I’m pretty comfortable with text/LLM prompting, but I don’t want to give him a vague or over-technical answer.

For those of you who already use Gemini for visual generation (images, video clips, animations, etc.), what would you tell him?

I’m mainly looking for:

• Real best practices that actually improve output quality
• A clear structure of what a good visual prompt looks like
• Common mistakes to avoid
• And maybe a couple of “perfect prompt” examples

Thanks in advance! 🙏