Why We Still Get on the Plane

In March we shot a brand film for Møller on the Danish coast. Two days, small crew, mostly natural light. On the second morning the fog came in off the water, unplanned, and gave us a shot that now opens the film: a craftsman carrying an oak chair down to the shoreline, half dissolved in gray. It tested better than anything else in the edit. Everyone who sees it asks about it.

You cannot prompt for that shot. I know, because half my industry spent this year trying to convince clients otherwise.

What the pitch decks get wrong

The pitch is seductive: why fly a crew anywhere when generative video can conjure any location, any light, any weather? The cost line goes down, the options go up. I understand why procurement loves it.

But the pitch misunderstands what a shoot actually produces. A location shoot is not a rendering job. It is a machine for capturing accidents. The fog. The way a real person's hands actually hold a tool they have used for twenty years. The line a founder says between takes that becomes the whole campaign. Every producer has a version of this story, because the unplanned moment is not a nice bonus of production. Statistically, it is the product. Ask editors where the best moments in their last ten projects came from. It was almost never the storyboard.

Generated footage, by construction, contains only what was asked for. It is a medium with no accidents. That makes it the perfect tool for exactly the shots where nothing true needs to happen, and the wrong tool for the moments that make people feel something.

Where the new tools earn their place

I am not a purist, and our production pipeline this year uses more AI than most of the people dunking on AI. Where it genuinely works:

• Previs and pitch. We can show a client the film before we bid it. This has shortened approvals by weeks and killed bad ideas cheaply.
• Plate extension and cleanup. Removing a crew shadow or extending a sky used to be a VFX line item. Now it is an afternoon.
• Versioning. Cutting one hero film into thirty formats and lengths is finally not a soul-destroying job.
• Scouting. Testing how a location might read in different seasons before committing travel budget.

The pattern: AI is superb around the shoot. The shoot itself is where the truth gets captured, and truth is the only thing audiences are rewarding this year. The backlash against synthetic advertising has been the loudest consumer signal of 2025, and it maps exactly to what we see in testing: audiences may not consciously spot generated footage, but their trust scores move like they can.

The honest math

Here is what surprised our clients this year. Real production got cheaper too. Smaller cameras, leaner crews, AI handling post-production drudgery. The Møller film cost roughly sixty percent of what an equivalent shoot would have cost in 2022, and looked better. The choice was never between an expensive real film and a cheap generated one. It is between two affordable options, one of which can contain fog.

So our recommendation to brands is not romantic, it is tactical. Spend generated pixels on the disposable middle of your content pyramid, the resizes and the product spins. Spend real days in real places on the work that carries your brand's emotional load: the anthem film, the founder story, the craft documentary. That work is your proof of life.

The plane ticket is not overhead. The plane ticket is the strategy.