Concept Exploration: RIFT

At Promise, every GenAI director is asked to bring forward ideas for films and series. I wasn’t new to writing scripts or concepts, but it was my first turn at bat at Promise — I wanted to swing for the fences.

Among the ideas I put forward was a film called Rift. This was the first thing I wrote down:

A cataclysm splits the Earth in two.

One half drifts into orbit around a rogue star and freezes. Decades later, survivors launch an expedition, only to find life has changed in impossible ways.

There was a lot to explore and questions to answer. How would anyone survive? How would the two halves of the world adapt in different conditions?

I gave the development team a five-sentence description of the idea, and they were intrigued, so they moved me to the next stage: a concept deck.

World-building with AI

First, I needed to figure out how anything could survive a planetary split. I’m not a scientist, and this is where AI proves very useful. I pressure-tested the science in ChatGPT for 2 days. 

I landed on something both plausible and interesting:

• One side stays in orbit—overheated, irradiated, everyone forced underground. No more government, space program, or supply chain.

• The other side freezes, where life undergoes rapid evolution (with a plot twist behind it).

I built mood boards and shared them with the team, testing if this was too ambitious, too complicated, or just maybe viable. I hadn’t lost them yet.

Mood Boards

Deciding on the Story

The concept itself was baked in both practical and emotional circumstances. People lost loved ones in the split and would want closure. With communication wiped out, decades passed with no idea what was happening on the other half of Earth—just the strangeness of seeing it hanging in the sky.

I expanded on the expeditioners: how pre- and post-split generations might clash culturally, and how people would view science after it failed to predict half the planet tearing away. I shaped it into a three-act pitch deck.

But the development team wasn’t as excited as I was. My first draft leaned too much on explanation and not enough on clarity. They wanted something more clear, and not overcomplicated.

I was Venn diagramming instead of keeping it simple:

Simplifying the Deck

So I rebuilt it. Fewer slides, tighter sentences. Just enough detail to sketch each act and introduce the main characters.

The new deck helped, but the team's next test was bigger: a concept trailer.

Cutting the Trailer

At Promise, a key part of the development process is to show what a story will look like early, as a kind of gut check: can the idea sell itself, does the story come through clearly, does it feel make-able? 

Normally a trailer is the last thing you make after the project is filmed; here, it’s the first. 

I drafted a trailer outline that would build intrigue without spelling out the whole plot. Then I pulled together the tools:

• MidJourney and Reve for images

• Magnific for upscaling

• Luma Dream Machine, Kling, and Veo 2 for animation

• Udio for score; Epidemic Sound and ElevenLabs for SFX

• ElevenLabs for voices (including my own voice acting)

• Video editing to pull it all together

  
  

It was important to me that the AI imagery didn’t feel derivative (hard to pull off). My AI skills were important: Rift wouldn’t work without believable visuals, so I put a lot of effort behind the AI VFX—cataclysmic explosions, wide shots of Earth from space, and two distinct versions of the same planet.

The first cut missed the mark. The team felt like they just watched a generic disaster movie. I focused on the visuals of the Earth splitting apart instead of the human story.

Iterating on the Trailer

So I went back in. Got feedback. And with notes from Dave Clark and MetaPuppet — two of the best AI filmmakers working today — I kept recutting. Slowly the trailer shifted from spectacle to story.

This is what I shared — not a finished trailer, but a proof of concept made to communicate the vision internally:

Concept Exploration: RIFT

While they loved the direction of the visuals, we all agreed that I needed to continue to work on the story and character development.

The Outcome

The process pushed me to refine ideas more precisely, pitch with greater clarity, and wield AI as a creative partner instead of a gimmick. Above all, it underscored the importance of story.

Since then, I’ve applied those lessons to new projects at Promise. My next story has already taken shape as a script for a forthcoming series — and that progress stems from what I learned building Rift.

It was my first time creating with Promise. It won’t be my last.

-- Dale Williams a.k.a. @TheReelRobot, GenAI Director at Promise