Can we talk about how good AI has become for creating product/startup videos?
I recently launched a campaign to raise some funds, and one of the questions I got was:
"Which agency did you hire to make the product video?"
The answer was: none.
The entire video was created in-house using AI tools. It took some experimentation, but the final result was good enough that people assumed it came from a professional production studio.
The first video got a lot of positive feedback, so I decided to use a similar AI workflow for my latest July 2026 iOS feature release.
This is the outcome: https://www.youtube.com/watch?v=re4qDCFjx70
For anyone curious, here's the workflow I used:
MindNote (of course)/ ChatGPT / Gemini – Brainstorming ideas, writing the script, and creating the storyboard.
Magnific – Defining the visual style, generating characters/models, and refining the overall aesthetic.
Gemini – Creating poses, generating product mockups, and maintaining visual consistency between scenes.
Hailuo AI – Animating the images.
Canva – Editing and assembling the final video + adding subtitles.
Uppbeat – Choosing the background music.
What surprised me most wasn't that AI could make a video it was how accessible the entire process has become. A year or two ago, I probably would've needed a designer, an animator, models, a budget for a photoshoot, a weekend recording session, and a video editor to produce something similar.
Having worked at large corporations and unicorn startups in the past, I know this is the kind of production that would typically involve multiple teams, weeks of coordination, and a significant budget. Today, a small team or even a solo founder can create something surprisingly polished with the right AI workflow.
I'd love to see what workflows everyone else is using. Any tools you've found that work particularly well for product videos? I think that Higgsfield is also awesome.
Strong workflow. One thing I’d add for founders using this kind of AI video stack: track cost per usable iteration, not just cost per final asset.
The real win is being able to test 5 hooks, 3 storyboards, and 2 CTAs quickly. But once you’re bouncing between ChatGPT, Gemini, Hailuo, image tools, and editing tools, it gets easy to lose sight of which part of the workflow is actually producing usable output vs burning tokens/credits.
For my own AI-heavy work I’ve found a simple rule useful: every video concept should have a rough “idea to first usable draft” cost and time. That makes the workflow repeatable instead of just impressive once.
Tiny relevant plug since you asked about tools: I’m building TokenBar for this exact kind of usage visibility on Mac, free to try: https://tokenbar.site/
Cool!
The biggest shift isn't that AI makes production cheaper.
It's that it changes the speed of iteration. When founders can create a new product video in hours instead of weeks, they can test different narratives just as quickly as they test product features. That feels like the more important change.