Not really, but I've been down that rabbit hole enough to know it's more hype than substance.
The real work is the same as any video production. You still need a script, a storyboard, a clear audience. The AI doesn't do that for you-it just adds more fiddly steps. People think they're going to skip the tedious part, but you end up fighting prompt limits, inconsistent outputs, and hallucinations that ruin the tone.
I've used Claude for script structure, OpenArt for reference images, ElevenLabs for voiceovers, RunWay for generation, and CapCut for assembly. It's a patchwork of tools that barely talk to each other. The costs? Who tracks that when you're chasing a half-decent result.
The real gotcha is that every AI tool has a ceiling-weird character consistency, audio sync drift, can't handle nuanced documentary pacing. If you want a slick, polished doc, you're still better off with a camera and a human editor. The AI stuff is fine for B-roll experiments, not for a finished product.
