AI video for beginners: prompts, workflows and references
Video by Dan Kieft
Key takeaways
- Describe a shot clearly: subject, action, camera and mood
- Use a repeatable workflow: image first, then animate it
- Reference images and videos keep your look consistent
- Starting from a strong still image is easier and cheaper
Dan Kieft's 27-minute beginner guide is the deepest AI video lesson in this path. It takes everything you need to start creating AI videos and puts it in one place: prompting, workflows, and using reference images and videos to control what the AI makes.
What it covers
The guide starts with prompting for video: how to describe a shot so the AI understands the subject, the action, the camera and the mood. Then it moves into workflows, the order in which you create images, turn them into video and combine shots, so you're not starting from zero every time. A big part of the video is about references: using AI images and existing videos as a guide, which is the most reliable way to keep a consistent look and get the motion you want. By the end, you'll know exactly how to start on your own.
Who it's for
Beginners who watched the shorter AI video lesson before this one and want the full picture. It's long, so watch it in parts and try each idea as you go. Dan also runs a free community where he shares his prompts.
Good to know
The workflow uses Higgsfield, linked in the description. The skills (describing shots, building from images, using references) carry over to any AI video generator.
Tip: start most videos from a strong still image. Generate or pick the perfect first frame, then animate it. It's easier to control and cheaper than generating video directly from text, and it gives every clip in your project the same look. Keep a folder of your best reference images to reuse.
Tools mentioned
Higgsfield AI is a creative suite for generating images, videos, and voice content from text prompts or reference material. It offers tools for editing and upscaling media, plus an AI agent that automates creative workflows. The product is available on both web and mobile, and includes an MCP integration for use within Claude, along with an API for developers. It's aimed at creators, marketers, and teams who want to produce video, image, and audio content, including ads and cinematic scenes, without traditional production tools.
You finished this path
Keep learning in the Academy, or find the right AI tools for your next project.