The AI-Accelerated Pre-Vis: From Storyboard to Animatic in 60 Minutes
From Napkin Sketch to Moving Picture: Your AI-Powered Pre-Visualization Lab
Will AI replace directors, animators, and cinematographers? Let’s get this out of the way: No. But as of July 11, 2025, a creative who knows how to direct AI will have an almost unfair advantage. The grueling, expensive, and time-consuming process of pre-visualization—turning a script into a tangible storyboard and a rough animatic—used to be the exclusive domain of studios with serious budgets. For the indie creator, it was a dream. Today, that changes. Forget the existential dread. Think of AI as your new, fully-staffed, in-house concept art department. One that works 24/7, never needs a coffee break, and can render a thousand different worlds before lunch. In this lab session, we’re not just ‘making pictures,’ we’re building a professional pre-vis pipeline, going from a blank page to a moving animatic in about an hour.
Welcome, fellow creator. My work lives at the intersection of code and canvas, algorithm and artistry. I’ve spent years exploring how these new tools can supercharge, not supplant, our creative instincts. What we’re about to do is a workflow I’ve honed for my own projects, one that bridges the gap between raw imagination and tangible, pitch-able content. We’ll be using two primary tools: Midjourney for generating world-class, stylistically consistent storyboard panels, and Runway for breathing life into them, creating a motion animatic that captures the timing and feel of a scene. This isn’t about pushing a button and getting a finished film; it’s about accelerating creative decision-making on an unprecedented scale.
Part 1: Establishing The Vision – The Consistent Character & World
The single biggest mistake new AI artists make is generating a series of disconnected, beautiful-but-incoherent images. A film needs consistency. The audience needs to recognize a character from a wide shot to a close-up. Before we can storyboard, we must first cast our lead and build our set. In AI terms, this means creating a strong ‘character prompt’ and a ‘style prompt’ that we will reuse constantly.
Our mission is to create a ‘Character Sheet’ first. This gives us a visual anchor. We will then use Midjourney’s new Character Reference feature (`–cref`) to ensure our protagonist looks the same in every shot.
The Prompting Studio: Character Sheet Generation
Open Midjourney. Our goal is to create a tough but weary private eye for a cyberpunk noir film. We’ll specify a ‘character sheet’ layout to get multiple angles.
Copy and paste this prompt:
/imagine prompt: character sheet for a male cyberpunk private detective, grizzled face, trench coat, neon-lit alley background, expressive poses, cinematic, consistent character design, moody lighting, style of a Blade Runner concept art –ar 16:9 –style raw –v 6.0
Once you get a result you like, ‘Upscale’ your favorite image. Now, right-click on the upscaled image, select ‘Copy Link’, and save that URL. This will be your Character Reference.
Strategist’s Log (Deconstructing the Character Prompt): ‘Character sheet’ and ‘expressive poses’ explicitly tell the AI our goal. ‘Grizzled face, trench coat’ provides the archetype. But the magic lies in the style directive: ‘style of a Blade Runner concept art’. This gives Midjourney a massive library of visual data to pull from, ensuring a cohesive and professional look. We use `–ar 16:9` from the very beginning because we’re making a film, not a square social media post. Always think about your final aspect ratio.
Part 2: The Storyboard – Building The Scene, Shot by Shot
Now that we have our ‘actor’ (the character reference URL), we can start directing. Let’s imagine a simple three-shot scene from our script:
- SHOT 1 (EST): Wide shot of our detective entering a grimy, rain-slicked noodle bar in the cyberpunk city.
- SHOT 2 (MS): Medium shot of the detective sitting at the counter, looking wary.
- SHOT 3 (CU): Close-up on his hand, hesitating over a datapad on the counter.
Using our Character Reference URL and a consistent style prompt, we will now generate these exact shots. This is where the human filmmaker’s knowledge of cinematography becomes the critical skill.
The Prompting Studio: AI Storyboarding
For each shot, you will use the same base prompt structure but change the cinematographic direction.
Paste your Character Reference Image URL first, then add the prompt.
SHOT 1 PROMPT:
/imagine prompt: [PASTE YOUR IMAGE URL HERE] cinematic wide establishing shot, a detective enters a grimy neon-lit noodle bar, rain slicked streets, cyberpunk city, moody, anamorphic lens flare –ar 16:9 –cw 100 –style raw –v 6.0
SHOT 2 PROMPT:
/imagine prompt: [PASTE YOUR IMAGE URL HERE] cinematic medium shot, a detective sits at the counter of a noodle bar, looking wary, background shows steam and holographic menus –ar 16:9 –cw 100 –style raw –v 6.0
SHOT 3 PROMPT:
/imagine prompt: [PASTE YOUR IMAGE URL HERE] cinematic extreme close-up, a detective’s hand hesitates over a glowing datapad on a wet counter, shallow depth of field, film grain –ar 16:9 –cw 100 –style raw –v 6.0
Strategist’s Log (Directing the AI): The `[PASTE YOUR IMAGE URL HERE]` is the `–cref` command in action. We are telling Midjourney, “Make this shot, and make sure the main character looks like this person.” The parameter `–cw 100` sets the Character Weight to maximum, ensuring the likeness is prioritized. Notice the cinematographic language: ‘wide establishing shot’, ‘medium shot’, ‘extreme close-up’, ‘anamorphic lens flare’, ‘shallow depth of field’. You are no longer a prompter; you are a Director of Photography, and these are your lenses and lighting instructions for the AI.
Part 3: First Movement – Animating The Storyboards with Runway
Still images are great for a printed storyboard, but an animatic helps us understand the timing and motion of a scene. This is where we bring our panels into our second tool, Runway. Using its Gen-2 model, we can feed it our static storyboard images and generate short (4-second) video clips, essentially breathing life into each panel.
This is not about generating the final VFX. It’s about getting a feel for pace. Does a slow push-in on the close-up feel more tense? Should the character enter the frame more quickly? These are the questions we can now answer in minutes, not days.
The Prompting Studio: Image-to-Video Animatic
Navigate to Runway and select the ‘Image to Video’ generator.
1. Upload Image: Upload your first storyboard panel (the Wide Shot).
2. Control Motion (Optional but Recommended): Click on ‘Motion Brush’. Paint over the character and the door. Set the directional controls to create a slow forward and right motion, simulating him walking in. Or, you can just use a text prompt.
3. Use a Prompt for Motion: In the prompt box, you can describe the motion: “slow push in, rain falling down, character walks slowly into the bar”
4. Generate: Click Generate. Runway will produce a 4-second clip. You can extend it to 16 seconds if needed. Repeat this process for each of your three storyboard panels, thinking carefully about what needs to move in each shot.
Strategist’s Log (Controlling Chaos): AI video can be unpredictable. The key is to constrain it. Using the Motion Brush in Runway is the single most powerful feature for pre-vis. Instead of the AI guessing what should move, you are telling it: ‘Animate this object, in this direction’. For the close-up of the hand, maybe you don’t use Motion Brush at all, allowing Runway to just add a subtle ‘breathing’ motion and flickering light to the shot. The less you ask for, the more stable the result. Remember: this is for timing, not for final pixel-perfect output.
Part 4: The Human in the Loop – The Edit
You now have a series of short video clips. The final step of the AI-augmented workflow is to return to a traditional creative environment. Import these clips into your video editor of choice—Adobe Premiere Pro, DaVinci Resolve, Final Cut Pro. Place them on the timeline, trim them to pace, add a temp music track, and drop in some simple sound effects (rain, a door chime, a digital hum).
In these last 15 minutes of work, the project transforms. It goes from a collection of AI-generated assets to your scene. You control the rhythm, the emotional beats, the audio-visual synthesis. The AI was your tireless production crew; you are still the director.
The Big Questions: Your AI Debrief
“What about copyright and using this for a commercial project?”
This is the most critical question. As of mid-2025, the legal landscape is evolving. Here’s the professional workflow: The AI-generated output (the animatic) is an internal pre-visualization tool. It’s for you, your team, and your pitch deck to investors. The final film would be shot with real actors and locations, using the AI-animatic as a precise blueprint. For most platforms like Midjourney, you own the assets you create under their Pro/commercial plans, but you should always consult their latest Terms of Service. Think of it as a concept artist’s sketch, not the final oil painting.
“How do I avoid that wobbly, ‘AI video’ look?”
First, embrace its purpose. Animatics are supposed to be rough. Don’t aim for perfection. Second, control the motion. The more subtle the movement, the less ‘morphing’ and instability you’ll see. A slow camera dolly (`-camera zoom in` in Runway’s prompt) is often more stable than complex character motion. Third, use a lower ‘Motion’ value (in Runway’s settings) to keep the original image more grounded. The goal is to convey intent, not to create a flawless final product from the AI.
“This sounds expensive. What’s a realistic budget for this workflow?”
Let’s break it down. A Midjourney Pro Plan is roughly $60/month, giving you the fast generation hours and commercial terms you need. A Runway Pro Plan is around $35/month. So, for under $100 a month, you have an infinite pre-visualization studio. Compare that to hiring one concept artist for one day. This isn’t a free hobby; it’s a professional investment that dramatically reduces traditional production costs. You can often get the work for a whole project done in a single month’s subscription.
Your Creative Sandbox Assignment
Your mission, should you choose to accept it, is to pre-visualize a simple, 3-shot sequence. Pick an idea: “An astronaut discovers a glowing plant on a desolate Mars,” or “A cat knocks a precious vase off a mantelpiece in a Victorian mansion.”
- Character/Subject: Generate a ‘character sheet’ for your astronaut or your cat in Midjourney. Copy the image URL.
- Storyboard: Write prompts for an establishing shot, a medium shot, and a close-up, using your Character Reference (`–cref`) in each one.
- Animate: Take just one of those shots—the most important one—into Runway. Use the Motion Brush or a simple prompt to create a 4-second clip.
Don’t spend more than an hour on it. The goal is to experience the speed of the workflow and see how quickly an idea can become a tangible, moving image.
Your AI Integration Plan This Week
- Monday: Idea Generation. Don’t touch the AI yet. Just write down three simple scene ideas on paper. Pick your favorite. Write a 3-shot-list for it.
- Wednesday: Midjourney Day. Spend 30 minutes generating the character sheet and the three storyboard panels for your chosen scene. Focus on consistency.
- Friday: Runway & Edit Day. Spend 30 minutes animating your three panels in Runway. Pull the resulting .mp4 files into a free video editor (like DaVinci Resolve) and add some free temp music.
- Sunday: Review. Watch your 15-second animatic. What worked? What didn’t? How would you direct the AI differently next time? You’ve just completed a full pre-vis cycle that would have taken days or weeks a few years ago. Now, what will you build next?
Post Comment
You must be logged in to post a comment.