Want to post videos but don't like being on camera? AI-generated videos might be the solution.
If you don't know what these are, they're videos made by platforms like Sora-2 or VEO-3. They look like you're talking into your phone, but they're created by AI.
Are they perfect? No. But luckily, there are people who:
Still don't know it's AI
Know it's AI but don't care (like me with some YouTubers)
If your audience is one of these types, you're good to go.
Second, there's a popular saying:
"AI will never be worse than it is today."
While some people who don't like GPT-5 might disagree, AI models generally keep getting better.
I'm saying this because even if you're not happy with AI video quality right now, we're probably just 1-2 updates away from AI videos looking exactly like real videos.
This has many big effects, and we don't have time to talk about them all today.
But what I can tell you is that making AI videos is a skill.
If AI videos start looking completely real in the future, the people who learned how to use today's basic models will have the biggest advantage.
They'll know how to market themselves and build businesses while everyone else is still learning.
And while people might hate on it for being slop today, I think we're 12-18 months away from AI videos looking completely real.
So where should you start?
Here's a basic guide for learning how to make AI videos.
AI UGC Mini Roadmap
While there are paid courses that teach this, here's a basic roadmap:
Decide what your video is about and why you're making it
Plan your scenes (if you need more than one)
Think about the big details (location, mood, clothing, etc.)
Use AI to create an image of you in that setting
Make 1-4 more images for each new scene (maximum 5 scenes)
Use AI to write your script for the first scene
Edit the script to make sure it's correct and sounds natural (not like AI wrote it)
Use AI to create instructions for the video
Put the script and instructions together (if not already done in Step 8)
Choose how you'll do the voice: AI voice, record yourself, or use an AI copy of your voice
Send everything to your AI video maker
Watch the video it creates
If it's good, do the same for all your scenes
If it's bad, see if you can use extra footage to hide the problems
If that doesn't work, ask the AI to make the scene again
Once you have all your scene videos, start putting them together
Finish your video and post it
As you can see, this takes time and multiple tries.
Even though this list covers everything, just one step - like asking AI to remake videos you don't like - can take hours.
Yes, some people are trying to automate this using n8n. But there are problems:
First, without a human checking quality, automation creates a lot of low-quality, useless videos.
Second, if you add a human to check the videos, they'd be overwhelmed with hundreds of videos to review and edit.
But that's not the main point here.
To actually follow this roadmap, there are many AI tools and platforms you can use.
To keep it simple, here's where we recommend starting:
Planning your scenes: Canva (has templates), or just write it down on paper
Planning the big details: Write it down or use a UGC prompt generator tool
Creating an AI image of yourself: You'll need to upload a photo of yourself and ask the AI to put you in the scene you planned.
Writing your scripts with AI: Depending on your experience, use ChatGPT or Claude, or try a UGC script writing tool
Creating video instructions: I have a two-step method that involves defining all the small details, then using a video generation tool. But to make it easier, you can use one tool for everything or just do it in ChatGPT or Claude.
Choosing the audio: If you're just starting, upload your script and let AI create the audio. Adding your own audio can be complicated.
Making the video: Keep it simple: Use an AI platform so you can access the better models that create 1080p videos (instead of the Sora app, which only makes 720p vertical videos). Later, you can try specialized UGC AI tools like Arcads.
Creating extra footage (B-Roll): Same as above. If you're serious about this, try AI video platforms.
Editing the video: Unless you know how to edit videos, you can do basic editing yourself on CapCut, or pay someone on Fiverr or Upwork to do it.
This is obviously a short summary. If you want to learn more, there are hundreds of tutorial videos on YouTube that explain all of this.
Yes, this takes a lot of work.
But so does doing your hair and makeup, getting on camera, and filming videos in a clean, well-lit place.
So in my opinion, automating AI video creation - using templates, tools, and software - is a much better option.
Also, just because these AI tools don't make "perfect" videos today doesn't mean they won't be able to next year.
By learning these skills now, while the general public ignores them, you'll put yourself in the best position possible to profit from AI video in the future.
Catch you next time
