Real where it matters, AI where it opens up possibilities: the case for hybrid video production

A recent campaign for Dutch telecom brand Odido, directed by Michel Mölder, has been circulating in production circles for a good reason. The approach is straightforward to describe but harder to execute well: shoot real actors in real setups, then replace almost everything around them using AI-generated environments. New locations, new atmospheres, entirely different worlds — built after the fact, around footage that was captured on set with care.

What strikes us about this methodology is not the technology itself. It is the discipline behind the creative split. The team made a deliberate choice about what had to be real: casting, performance, styling, the small unrepeatable things that happen in front of a camera. And they made an equally deliberate choice about what AI could do better: scale, variety, visual ambition that a location budget would never stretch to cover.

That distinction matters enormously for communication teams in the EU bubble, where video briefs often collide with tight budgets and ambitious scope. A policy campaign needs to feel credible and human — but it also needs to land across multiple contexts, languages, sometimes entire different visual registers. Hybrid production is one honest answer to that tension. You invest in the human core. You use AI to multiply what surrounds it.

At White Rabbit, this is increasingly how we think about video work for associations, institutions and advocacy organisations. The production logic is changing fast — tools evolve almost week by week, as Mölder himself notes — but the underlying principle is stabilising: protect what only a real shoot can give you, and use generative tools to expand what would otherwise be out of reach. That is not a shortcut. It is a more honest allocation of creative effort and budget.

If your team is planning a video project and wondering how far a production budget can realistically go in 2025, hybrid methodology is worth putting on the table early. The conversation is less about AI versus traditional production and more about where each approach earns its place in the same piece of work.