Post-Production After AI

How AI is reshaping editing, storytelling, and creative decision-making

AI has quickly become one of the most disruptive forces in the creative industry. As companies downsize and markets tighten, many creatives are asking the same question: will AI replace our jobs?

From the perspective of an editor and post-producer, the answer is more nuanced. AI is undeniably changing post-production workflows, but it hasn’t replaced creative judgment. It has simply shifted where value lives.

Image and video generation tools are advancing at a rapid pace. Platforms like Midjourney, Veo, and Higgsfield are pushing visual quality further with every release, even if they’re not yet consistently broadcast-ready. Voice generation tools such as ElevenLabs have already reshaped traditional voiceover workflows, now commonly used across pitches, manifestos, internal films, and social content. Music generation platforms like Suno are also gaining traction as stock libraries increasingly feel repetitive.

Yet despite all this progress, storytelling still lives in the edit.

I haven’t encountered a tool that can autonomously shape narrative, pacing, emotion, and intent with genuine creative autonomy. AI can generate assets at scale, but it lacks taste. And taste remains the differentiator.

From a post-production standpoint, orchestration is everything. Knowing which tool to use, when to use it, and why is what separates a considered AI-assisted film from something that feels generic or premature. A line I recently came across summed it up well: AI doesn’t replace talent, it exposes taste.

There’s also a practical reality to this shift. Most AI tools operate on subscription models, with new platforms and updates appearing constantly. Subscribing to every new model isn’t sustainable. The real skill lies in understanding what can be achieved with freely available tools, where paid solutions genuinely add value, and how to combine them efficiently without compromising quality or timelines.

To move beyond theory, we recently tested a full end-to-end AI-assisted production. The goal was to create a complete 30-second perfume advertisement. For the purpose of the experiment, we chose a luxury reference brand: Dior. Credit goes to Mahmoud Elsamman, who led the generative side of the process and committed fully to exploring the toolset.

Art direction was developed using Midjourney. Moodboards and early visual exploration were created in Google Labs’ Mixboard, which also helped refine prompts and establish a consistent visual language. The script was generated using Gemini, while voiceover and music were produced through ElevenLabs.

For video generation, Mahmoud opted for Freepik’s node-based system, specifically for its ability to maintain visual consistency across shots. Through continuous iteration and prompt refinement, the model was effectively trained to sustain a cohesive look and feel, making it possible to generate additional scenes while preserving continuity.

Once the scenes were generated, they were passed into post. While the resolution was limited to roughly half-HD, the results were impressive given the speed and flexibility of the process. Editing and sound design were completed in Premiere Pro, resulting in a fully realised 30-second film.

The final output wasn’t without imperfections. Artefacts and inconsistencies remained, and human judgment was still required throughout the edit. What stood out most, however, was how fluid iteration had become. At one point, it became clear that the film was missing a transitional shot. After describing the scene, a new shot was generated within minutes, thanks to the previously trained model.

That moment felt like a glimpse of what’s coming.

This shift is no longer theoretical. AI has already integrated itself into traditional creative workflows, quietly and irreversibly. The opportunity now is not to resist it, but to understand it, shape it, and use it with intention. For editors and post-producers in particular, the role is evolving from execution to orchestration.

AI won’t replace the edit. But it will change who controls it.