The All-In team recently asked me to create an intro video for their holiday party with Tony Hinchcliffe. The idea was simple: show the audience how the besties actually got to the show. It was a fun opportunity to use a full AI pipeline to contrast the different travel styles of the group. While the video shows Chamath arriving from an Italian villa on a luxury yacht, it highlights a much different journey for Jason.
Below is the 6-step workflow I used to bring this project to life.
Full Holiday Party Intro Video
Six Steps:
1. Scripting
The process started with a few minutes of brainstorming to build out the narrative. We wanted to show Jason landing at a public airport, waiting for his own bags, and driving his own Uber X to the event.
Figma Page: I created a Figma page to track the steps and organize assets.
Imagery Planning: I wrote the script specifically to include visual moments that would work well as AI prompts later.
2. Voiceover Generation
I found that ElevenLabs was the best tool for this specific narration.
The Voice: I selected “Brian” for the project.
Style Exaggeration: I used the style exaggeration slider to add more personality and punch to the voice.
Audio Polishing: I generated multiple takes with different slider settings and used CapCut to splice the best parts into one final track.
3. Asset Collection
Before jumping into generation, I gathered organic assets from the internet. This included real photos of Jason and reference shots of airport scenes. Keeping these organized in Figma helped maintain the visual logic of the story.
4. AI Image Generation
I used Higgsfield to create the final images that would eventually become video clips.
Character Consistency: I dragged real photos of Jason into the tool and used prompts to place him into the new scenes.
The Uber X Detail: To make the branding look right, I manually placed an Uber logo on the car door image before running the final generation. This ensured the AI didn’t hallucinate the text.
Specific Prompts: I was careful to reference specific images (e.g., “The man in image 2 in the scene of image 1”) to get the best output.
5. AI Video Generation
Once the stills were set, I moved to the Create Video section in Higgsfield to add motion.
Settings: I used VO 3.1 and kept the clips to 6 seconds for a tight pace.
Directing the Motion: I used simple prompts like “man walks past the camera” or “getting into the car and pulling away” to keep the movement realistic.
Whisper Flow: I utilized Whisper Flow for voice-to-text prompting on the more complex driving shots.
6. Video Composition
The final step was the assembly in CapCut.
Syncing: I imported all the generated video clips and synced them to the polished audio track.
Final Export: After adjusting the timing to match the narration, I exported the final teaser.
Thanks for watching!
🤖 If you want to stay ahead of the curve on all things AI, make sure to join our community across all platforms:
📩 Get the Weekly Newsletter: https://thisweekinai.ai/
📺 Subscribe on YouTube: www.youtube.com/@ThisWeekinAIPodcast
📸 Instagram: www.instagram.com/thisweekinaipodcast
📱 TikTok: www.tiktok.com/@thisweekinaipodcast
