Case study: Integrating AI in Production Pipeline
Sense
Framing Context
87 Seconds is seeking new ways to integrate AI into their production pipeline, focusing primarily on pre-production. The central objective is to automate and accelerate certain steps while maintaining high-quality results that align with the company’s mission: ‘Content that makes people feel’. This project explores how generative AI tools can support emotional impact in creative video production.
Framing Goals & Intent
Research Question:
How can AI tools be integrated into a production workflow to enhance the emotional quality and efficiency of creative content for a video agency?
The goal is to assess tools such as ChatGPT, Midjourney, DALL-E, Stable Diffusion and Feel-in AI in terms of usability, personalizability and emotional relevance. Key factors include dataset control, visual fidelity, storytelling capabilities, and workflow alignment with 87 Seconds.
Sample
Exploring Ideas
Research included:
– Reviewing academic and industry sources on AI in video pre-production.
– Comparing AI tools used in concept development, script writing, and moodboarding.
– Studying tools like Filmustage (script breakdown), Midjourney (image generation), DALL-E (concept visualization), and Feel-in AI (emotional testing).
– Testing prompt structures, embedding personalization in GPT, and assessing tool performance in emotional storytelling.
– Exploring integration points into 87 Seconds’ creative process.
Prototype
A speculative AI-enhanced workflow was created:
– Brief > Concept > Script > Moodboard > Animatic > Audience Testing
• Tools used: ChatGPT (brief analysis, scriptwriting), Midjourney (style and visuals), DALL-E (storyboards), Eleven Labs (voice-over), Feel-in (emotional validation).
• A GPT was trained as a ‘creative director’ for 87 Seconds.
• A proof of concept campaign was created for CEMEX Ventures, including a visual storyboard, moodboard, script and final concept titled ‘Futuristic Construction World’.
Shape
Curating and refining
Each AI tool was evaluated:
– ChatGPT (versions 3.5 & 4) proved highly effective for ideation and brief analysis, less so for real-world emotional nuance.
– DALL-E showed strong creative visualization but had stylistic limitations.
– Midjourney offered strong results in hyperrealism and was preferred for visual quality.
– Eleven Labs delivered realistic voice-overs tailored to tone.
– Feel-in was tested for analyzing viewer emotions during content playback.
Prototype
A speculative AI-enhanced workflow was created:
– Brief > Concept > Script > Moodboard > Animatic > Audience Testing
• Tools used: ChatGPT (brief analysis, scriptwriting), Midjourney (style and visuals), DALL-E (storyboards), Eleven Labs (voice-over), Feel-in (emotional validation).
• A GPT was trained as a ‘creative director’ for 87 Seconds.
• A proof of concept campaign was created for CEMEX Ventures, including a visual storyboard, moodboard, script and final concept titled ‘Futuristic Construction World’.
Stage
Integrating Result
The project delivered:
– A custom GPT trained for pre-production creative work.
– An AI-based pipeline proposal aligned with 87 Seconds’ workflow.
– A proof-of-concept video concept with creative visuals and narration.
– Recommendations for integrating and customizing AI tools based on client context.
This case study illustrates how AI, if carefully embedded, can amplify emotional storytelling without replacing the creative human at the center.