Prompting 101: Video
Generative AI Essentials · Hands-on micro-session
Session Objective
We took the image skills from last week and turned them into video. The key tool was Google Flow - specifically the Frames to Video mode, which lets you upload a reference image as your first frame and then prompt the motion from there. The main framework was VISION: Visual anchor, Intention, Sequence, Immersion, Orbit, and Narrative. Together, these six elements give you director-level control over short AI video clips without a production team or editing skills.
Your Challenge
Complete the following steps
STEP 1
Build your reference image (3 min)
Use any AI image tool (Google Labs, ChatGPT, Gemini, or similar). Apply the 5-element image framework: subject, style, setting, composition, details. Create a single image that could serve as the opening frame of your…
STEP 2
Write a VISION prompt (4 min)
Write a video prompt using the VISION framework. Your prompt must include all six elements: - V - Visual anchor: Describe the reference image and what to lock in - I - Intention: Why does this scene exist? What is it…
STEP 3
Generate in Google Flow (3 min)
Open lab.google.com/flow. Create a new project. Choose Frames to Video. Upload your image as the first frame. Paste your VISION prompt. Generate one or two variations. Export the clip.
What to Submit to Your AI Buddy
Your full VISION prompt (all six elements, labeled)
The name of the tool you used to create the reference image
A one-sentence description of the work scenario this video is for
One thing you would change about the output if you ran it again
Grading Criteria
All six VISION elements are
V, I, S, I, O, N each appear clearly in the prompt. Missing elements = incomplete submission.
The sequence has 3 or more d
The S element must describe specific actions that happen step by step, not a vague summary. "People talk" is not a beat. "The employee looks at the manager nervously, then the manager slides a document across the desk, then the employee rea
Camera language is specific
The orbit element uses real camera terms: pan, dolly-in, dolly-out, tilt up, tilt down, rack focus, slow zoom, static wide shot, tracking shot. "The camera moves" is not specific.
The scenario is grounded in
The intention element connects to an actual professional use case: a team announcement, a product demo, a client communication, a training clip, a weekly briefing. Personal or abstract scenarios do not qualify.
