Have you ever scrolled through social media and seen those insane AI-animated shorts that look like this?
I decided to try it out for myself by creating a short for my Robinhood Gold Card article. But also I didn't want to spend any money on making "AI slop" content like this. So here's the behind-the-scenes look at how I did it for free using multiple AI tools.
The AI Stack
For this project, I combined several cutting-edge AI technologies:
- Google Flow/Whisk: Workflow automation and orchestration
- ChatGPT Object talk MCP: Intelligent object manipulation and narration
The Process
1. Content Generation
First, I used ChatGPT to analyze the Robinhood Gold Card and generate a script. I basically just copied and pasted the script from my article and asked it to make it more concise and suitable for a video using the ChatGPT Object talk MCP.
Prompt: Pixar-style 3D animated character based on a premium gold credit card, standing upright as a living character on a minimalist desk. The card is anthropomorphic (slight thickness, soft rounded edges, friendly proportions). Face is integrated directly into the card surface (no floating head). Eyes: large, expressive, intelligent, with subtle shine. Eyebrows: flexible, expressive, slightly raised in a teaching/confident expression. Mouth: smooth, animated, calm smile while speaking. Arms: soft, simplified cartoon limbs (rounded, no fingers). One arm pointing outward, one arm gesturing like a teacher explaining a concept.
Design style: Smooth Pixar-quality materials (soft gold, slightly reflective but not realistic chrome). Subtle gradients and warm highlights. Slightly exaggerated proportions to enhance personality.
Camera: Slight low angle (makes character feel authoritative). Subtle depth and parallax between character and floating elements.
Animation cues: Character "talks" with subtle head nods and hand gestures.
Tone: Confident, intelligent, and slightly playful teacher persona. Feels like a mentor explaining a financial advantage.
Constraints: No real logos, no brand marks. No text distortion. No hyper-realistic textures (stay stylized and animated). Pronounces "Robinhood as "raah-buhn-hood"". Completes the entire script within the allotted timeframe.
Script: "I'm the Robin hood (pronounced raah-buhn-hood) Gold Card. I get 3% back. 3% IRA match. $1,000 free margin. Use me right and I pay you lots of money!" (make cash appear on table)
2. Visual Creation
Using Whisk, I generated an image using the ChatGPT Object talk MCP script as well as the actual card image as context.
3. Workflow Automation
After uploading the image to Google Flow, I then again copied and pasted the condensed script from ChatGPT Object talk MCP.
Technical Challenges
The generated videos had 2 main issues:
Vocal Accuracy: Ensuring the AI generated voice matched the intended tone and pace. In this specific case, the AI had trouble pronouncing "Robinhood" correctly.
Unwanted Visual Artifacts: I had to run the image generation a total of 10-15 times to get what I wanted without any unwanted elements or artifacts.
Credit Limitations: I wanted to test this without spending any money upfront. I was able to make as many images as I wanted with Nano Banana without any issue, but I was limited to about 5 videos (20 credits each, 100 credits per day) on the Veo 3.1 model at a time. I feel like companies may make this more restrictive in the future.
For both issues, I have to mess with the prompt and specifically call out exactly what was wrong and what I wanted to change for it to produce the correct result. I also had to just omit some features entirely because it couldn't produce it in enough iterations without hitting my token usage limit.
The Results
The final video demonstrates what's possible when you combine multiple AI tools. What would normally take hours of manual work was completed in minutes, with the ability to regenerate and refine instantly.
So I actually posted this as a YouTube Short, and here's what happened: it got 2.3K total views within 12 hours before the algorithm cut it off. The retention rate was a measly 24.7%, which honestly makes sense why YouTube stopped promoting it.
What's surprising though is that I didn't expect any views beyond the people I intentionally shared it with - this was my first YouTube Short ever. The fact that the algorithm gave it any traction at all was actually encouraging.
What I learned from this is that quality still matters despite the AI-animated novelty. Viewers want authentic, engaging content, and the first 3 seconds determine the outcome. I'd argue that once the novelty wears off, it might actually be a lot harder to get traction without a lot more effort as people stigmatize AI-generated content more and it becomes regulated.
If I were to do it again, I'd focus more on capturing more attention in the first 3 seconds and less on dumping as much info as I could within 8 seconds. The technology is impressive, but at the end of the day, you still need to monitor engagement tactics.
What's your biggest concern about AI-generated content like this?
The Bigger Picture
Whether we like it or not, these AI image and video tools are already producing examples of AI slop like "Fruit Love Island" that make you question if we've peaked as a society. But beyond the cringe, they also represent a new method of content creation.
This isn't just about making one video about a credit card. It's about demonstrating a new mode of content creation where:
- Ideas become reality almost instantly: Even though I took a whole afternoon to make this 1 video, the actual process of generating the content was incredibly fast. With future creations, I could generate content at a much faster pace.
- Complex projects become simple: All I needed was an idea, and after that AI tools can take care of the rest and help me refine it into reality.
- Scalability is real - and pioneers are using it for monetization: If I actually paid for these services, I could generate content at an unprecedented rate, opening the doors to monetization. I've seen relatively new social media accounts posting consistently and making steady income while capturing millions of views.
The tools I was using handled the entire technical execution while I focused on the creative direction and strategic refinements. For storytellers who want to create content faster and more efficiently, this can be a game-changer if used in a meaningful way. But if used in a way that doesn't add value, it can actually hurt your brand and add to the rising environmental concerns of AI usage.
Thanks for reading!
Like this kind of breakdown?
Sign up for updates — no spam, no affiliate fluff, just helpful notes like what you just read above.