Learn the 6-step AI video production workflow and produce a finished piece using it. Process over tools. Repeatable pipelines over lucky shots.
What you'll be able to do
Follow a repeatable 6-step workflow to produce AI video from scratch
Use node-based creative tools to build consistent, reusable generation pipelines
Add professional voiceover and sound design to any video project
Know exactly when to stop generating and start editing
The artifacts you'll build
The concept
Most people treat AI video like a slot machine. Type a prompt, hit generate, pray for something usable. When it looks weird (it will), they tweak random words and try again. Forty minutes later they've burned through credits and have nothing to show for it.
That's not a workflow. That's gambling.
The people who consistently produce good AI video follow a process. Six steps. Each one feeds the next. Skip a step and the whole thing falls apart. Do them in order and you'll produce better output in less time with fewer wasted generations.
Here's the workflow:
Step 1: Ideation. This is the hardest part and the one you can't automate. Your output is only as original as the idea behind it. To get great ideas you need to consume great content. Not AI content. Real content. Films with strong cinematography. Photography books. Music that hits different. Ads that stopped your scroll. For artistic projects, go to the source material that inspired entire movements.
For content that sells (AI influencers, product ads, social clips), stay plugged into what's working right now. Scroll with intention. Save what catches your eye. Inspiration isn't passive. It's a practice.
Step 2: Scene brainstorming. Once you have the idea, you need to describe what's happening in the video. Every detail. The characters, the setting, the movement, the mood, the camera angle, any dialogue or text.
This is where Claude or ChatGPT becomes useful. You describe the scene as best you can, and the AI helps you articulate it properly. It fills in the cinematographic vocabulary you might not have. "Woman walks through neon alley" becomes a fully fleshed scene with lighting direction, camera movement, focal depth, and atmosphere. The AI puts the right words on your vision.
Step 3: Character and element creation. You already learned this. In Module 7, you built a reference-first workflow for image generation: lock down your characters, environments, and visual elements as images before touching video.
That same principle applies here with one addition. You also create your starting frame and ending frame. These two images become the anchors for your video generation. The AI model knows where to begin and where to end, which gives you control over the output that text prompts alone can't match. Consistency across shots comes from image references, not from hoping the model remembers what your character looks like.
Step 4: Generation. Name every file clearly. Your character images, your start frame, your end frame. Write your prompt referencing those file names. Then generate. Right now the best platform for building these workflows is the Freepik AI Suite.
It's not just another generator. Freepik Spaces gives you a node-based canvas where you connect image generation, video generation, voice, sound effects, and upscaling in one visual pipeline. Each node does one job. You chain them together. Change one input and the whole workflow updates.
It's how serious creators build repeatable processes instead of one-off lucky shots. You get access to Veo 3, Kling 3.0, Runway, Seedance, and more, all in one interface. Switch models to find what works for your specific shot. Here's a full course on the Freepik workflow.
Step 5: Editing. Raw generated clips rarely go straight to publish. You'll cut, caption, arrange, and time them. CapCut handles this for most people. It's free, it's fast, and it does everything you need for social content. If you have editing chops and want more control, Premiere Pro or other tools are there. The point isn't which editor. The point is that editing is a separate step from generation, and treating it that way keeps your process clean.
Step 6: Audio and voiceover. This is where most AI video goes from "cool demo" to "actual content." ElevenLabs is the leader right now. 1,200+ voices, 29 languages, and the most natural emotional delivery on the market. Either way, the audio layer is what makes people watch past the first two seconds. Voiceover gives your video a heartbeat. Sound design gives it a world. Don't skip this.
That's the workflow. Six steps. The prompt in this module walks you through building and testing each one with your own project.
One more thing. AI video tools change fast. Models that lead today get replaced in weeks. But the workflow doesn't change. Ideation feeds brainstorming feeds references feeds generation feeds editing feeds audio. New model drops? Same steps. Better raw material. That's why we're teaching the process, not the buttons.
The prompt
The prompt turns the AI into a video production strategist who thinks in workflows, not features. Someone who walks you through building your complete pipeline step by step, testing each step before moving forward.
Before you paste
Make sure you're in your course project
Your my-context.md, my-learning-style.md, and my-knowledge.md files should be attached to the project
Have a video project in mind (something for your actual work, not a hypothetical)
After the exercise, run the Learning Extraction Prompt to update your knowledge file and save your session notes
What happens in each phase
PHASE 1 Context Bridge: determine what video content matters for your actual work
PHASE 2 Ideation & Scene Design: find inspiration, brainstorm a full scene with AI assistance
PHASE 3 Visual Pipeline: build references and design the generation workflow with nodes
PHASE 4 Polish Pipeline: choose editing tools, design your audio and voiceover layer
PHASE 5 Produce & Document: create one piece using the full pipeline, build your reusable workflow document
GATE Mastery Gate: 7 real-world scenarios testing the complete production workflow
What you should see: The AI reads your files, asks what kind of video you need for your work, then walks you through each step. You'll brainstorm a scene with AI assistance, plan your visual references (pulling from your Module 7 skills), design a generation workflow, choose your editing and audio approach, and map the whole thing into a reusable production document. By the end, you'll have produced one piece using the full pipeline. Expect 30-45 minutes of active work.
After this module
Run the Learning Extraction Prompt to update my-knowledge.md with what you learned.
Save to module-08-video-audio/
production-workflow.md: your personalized video production pipeline
Your finished video piece (or link to it)
session-notes.md from the Learning Extraction Prompt
Learning Extraction Prompt
After completing the module prompt above, paste this into the same conversation. The AI reviews everything that just happened and extracts what you actually learned, not what was presented, but what you demonstrated.