Use Case: Visual Dreamwork with AI — A New Creative Workflow

Overview

In this exploration, I leaned into the world of generative AI and cinematic rendering to create two dreamy, high-fidelity artworks that blur the line between imagination and photorealism. This use case showcases the output of a new visual workflow I picked up from an inspiring X thread by @BrettFromDJ. The goal? Test the boundaries of storytelling through image—using AI as a visual co-pilot.

The Process

Inspired by Brett’s step-by-step process, I followed a cinematic image creation flow that combined prompt engineering, post-processing, and stylized atmospheric effects. Here’s a high-level breakdown:

• Prompt Crafting: The workflow begins with thoughtful image prompts optimized for soft lighting, cinematic composition, and depth. I explored themes like “manifestation,” “dream states,” and “visual storytelling”—themes I often integrate into branding or conceptual work.

• Image Generation: Using Midjourney, I guided the AI toward stylized yet grounded visuals. I emphasized natural environments with surreal lighting and dreamlike architecture, leveraging terms like “volumetric light,” “C4D render,” and “8K composition” to push fidelity.

• Upscaling: I then take it into Leonardo or Giga Pixel (sometimes both) to upscale the image, add more pixel depth, and increase the detail.

• Post-Processing: After generation, I brought the visuals into Photoshop and Lightroom for color grading, contrast enhancement, and to further blend the natural and surreal elements. I also added the Path blur effect and erased the focus object (the tower and the tree/foreground) to make it clear this is where I want the viewers eyes to focus.

• Final Processing: Final processing occurs in Figma, and this is where a noise/texture effect is added, the typography, and my logo with website at the bottom. In addition, I created a dot matrix of the main object (in ChatGPT) and overlayed this via new layer and different blend modes like a screen or color dodge. Lastly, I added some drop shadow effects to the Japanese inspired image to help lift the text and make it more visible against the white background.

• The dot-matrices that were overlayed onto the finals:

  

This AI-assisted visual workflow involved a range of creative tools for different stages of the process:

• ChatGPT – Prompt-based image generation and text.

• Midjourney – Prompt-based image generation

• Leonardo AI – Creative generation and upscaling

• Gigapixel AI – Pixel-level upscaling and detail refinement

• Photoshop – Color grading, contrast, overlays, and final image polish

• Lightroom – Additional tonal correction and mood control

• Figma – Final composition, typography, layout, branding, and visual effects

Figma File
https://www.figma.com/design/64i8PjFn4yGAgzWy6Zrv6e/Mock-Preplexity-Ads?node-id=0-1&t=sOtOc7puuQui8S99-1

Bonus

My logo chrome and with a reflection of each image. Very cool!

One shot prompt with ChatGPT and adding my chrome logo then the landscape image and this prompt:

"Apply the image of the object in the left image to the reflection on chrome object on a right image. Maintain the exact shape and form of the object in the right image. Solid black background #000000. 8k resolution."

To easily change the reflection image, I then add this prompt and upload the next image in the same ChatGPT chat.

"Now change the image reflecting on the chrome to be this one I am uploading now"

Viola!