AI video is having a moment, but most marketers still treat it like a magic button. After watching Kieran Flanagan and Kipp Bodnar break down how they built a two-minute ad with VO3.1 and Nano Banana Pro, my take is simple: the idea still wins. Tools only shine when guided by taste, structure, and skill. That’s not a warning. It’s the advantage for creators willing to do the work.
What They Got Right
Kieran didn’t sell a fantasy. He put in 20–25 hours to get one video right, then showed how a repeatable process could cut the next run to a few hours. That honesty matters. He and Kipp focused on repeatability, not hype.
“You cannot outsource the idea… AI is a good thought partner, but it’s not going to give you an idea that’s better than people with real taste.”
That line nails it. AI accelerates craft, not vision. If you’re an expert, it multiplies you. If you’re stuck in the middle—half-skilled, half-confident—you’ll stall.
The Process That Actually Works
Their approach maps to how pros think, even if you’re new to video. It’s disciplined and simple.
- Start with a clear concept. Write it out. Don’t let a model invent it.
- Build a storyboard with scene, visual description, audio, and dialogue.
- Go “ingredients to video,” not text to video. Create reference images for consistency.
- Generate 8-second clips in VO3.1, then stitch in a basic editor like iMovie or CapCut.
This “ingredients first” move is key. Reference images keep characters, clothing, and rooms aligned across scenes. That’s how the ad kept its 1970s sitcom style without drifting into chaos.
“What the best AI video creators are doing is spending a lot of time on the look and feel and the base images… then the video output is much better.”
Where AI Still Trips
They ran into real friction. Dialogue between two characters broke. Names triggered copyright flags. Lip sync went wonky. Some scenes took hours to fix. Kieran even had to “kill” a character mid-scene to get the other voice to speak.
“Anytime I had dialogue, it would just have Teddy say the entire dialogue… So I told it to kill Teddy.”
That’s hilarious and instructive. AI is not your editor-in-chief. It’s your fast assistant with odd blind spots. It also shows how far $2 worth of generation can take you—cheap to iterate, expensive in time if you lack a plan.
My Take: Use AI, But Keep the Pen
I’ve built businesses across crypto, social, and online media. The lesson repeats: own the story, then scale the output. Their sitcom concept worked because it had a hook—“confidently wrong AI”—then used AI to flesh it out. That’s the order.
If I were advising a marketer today, I’d push this playbook:
- Pick a human insight first. “Confidently wrong” beats “AI can do video.”
- Storyboard with strict time boxes. VO3.1’s 8-second rhythm forces clean cuts.
- Lock references early. Characters, lighting, and set design must match.
- Plan workarounds. Use 11 Labs for voice. Expect name/copyright limits.
- Ship short. Make 30–60 second cuts for faster testing and better watch time.
One more point they hinted at: experts and true beginners both do well. Experts shape the output. Beginners try wild ideas without fear. The stuck middle wants the tool to do the thinking. It won’t.
Evidence That Should Sway You
Time savings were real. First attempt: roughly 20–25 hours. Next version with the process: a couple of hours. That’s an 80–90% reduction in time, with control over style and structure. Cost? Roughly $2 across image and video generations. For small teams, that’s game-changing.
“AI is incredible for people with real domain expertise… it’s going to be an accelerant.”
That’s the core signal. Treat AI as leverage on top of taste, not a substitute for it.
Final Thought
Great AI videos start with a great idea. Then they get built through a clear, repeatable process. If you’re serious about marketing, stop asking the model for magic. Write the concept, frame the scenes, control the look, and use AI to sprint.
Make one two-minute ad this week with the “ingredients to video” approach. Publish a 60-second cut. Watch how fast you learn. The winners will be the ones who keep the pen—and use the tools to move faster than everyone else.
