Good instinct to ask before building — you're already doing the thing you're selling.
On your questions:
1. Figma clickthroughs + a Loom walkthrough sent to 10-15 target users. Messy but cheap.
2. "Fake" video data is fine if you're measuring the right thing. Click-through on a landing page beats survey intent every time. People lie to be nice; clicks don't.
3. Honest answer: I'd pay if you could show me a case where the video said "no" and saved someone real money. One solid example beats a hundred testimonials.
Bigger concern — 2 devs saying they'd use it isn't validation, it's encouragement. Before you build anything, I'd run your own video prototype of this service and see if strangers convert. Meta, but you'd learn fast.
What's the actual customer you're picturing — solo devs, or teams with some budget?