Vinne Schifferstein Vidal, co-founder, MC&V
Over the past few months, there’s been a noticeable rise in production companies repositioning themselves as “AI-first.” Some are new creator-led entrants, while others are established agencies layering AI into existing production models.
We’re part of that shift ourselves. Earlier this year, we launched MC&V, an AI-native production company, after several years working inside this space, building workflows, delivering commercial projects and investing in R&D. That experience has taught us where AI holds up and where it doesn’t.
From that vantage point, one issue stands out clearly: AI production doesn’t have a tool problem. It has a producer problem.
Right now, the market is splitting into two camps, both struggling for opposite reasons.
On one side, you have creator-led shops. Talented people who understand the tools and can produce impressive outputs, but who often operate without a proper production layer. The assumption is that AI simplifies the process to the point where clients can work directly with creators.
In reality, that’s usually where things begin to break down.
Clients’ brief business objectives, brand nuances, stakeholder expectations, and evolving feedback. They don’t necessarily have the language or experience to brief a creative technologist, coder or AI specialist directly.
Translating those requirements into something AI systems can respond to while managing iteration, consistency and a defined outcome, is not simply a creative exercise. It’s a production discipline.
We’ve seen this firsthand on projects where a brief expands beyond a single output into multiple formats, markets and iterations. Without a structured production layer, even strong creative work starts to fragment. What initially feels fast and efficient quickly becomes harder to control as complexity increases.
On the other side, you have traditional agencies launching AI production capabilities under the guidance of conventional film producers. Experienced people, no doubt, but often operating from production models built for a completely different environment.
And that’s where it becomes a mismatch.
AI production is not a compressed version of a traditional shoot. It doesn’t follow neat linear stages or rely on fixed outputs. It’s iterative, system-driven, and heavily dependent on how workflows are designed up front.
Applying a traditional production mindset to that environment often creates inefficiencies. You end up forcing a non-linear process into a linear structure, and the results rarely hold up under the realities of commercial production.
If you look at more mature markets like the US and parts of Europe, the companies succeeding in this space are recognising that AI production requires a different kind of producer altogether. Someone who understands not just timelines and budgets, but systems, iteration and behaviour.
Someone who can translate between client expectations and machine outputs while managing consistency, IP, copyright, data and brand constraints responsibly.
This is another area where the experience gap is becoming increasingly visible. Conversations around ownership, ethics and provenance are still underdeveloped in generative AI, and very few producers have developed that expertise through real commercial experience.
We’ve gained experience in that area ourselves, learning by doing and working closely with experts, including Josephine Johnston, CEO of the Copyright Agency, to stay current as the landscape evolves.
Which is why so much of the work still feels inconsistent. You see strong individual outputs, but far less evidence of repeatability. You see experimentation, but less proof of production at scale.
We’ve seen this ourselves. Early AI work is often about proving what’s possible. But once you move into real production, with multiple assets, stakeholders, timelines and approvals, the question becomes whether you can actually deliver consistently. That’s where most projects either hold together or start to fall apart.
The differentiator, then, is not who has access to AI. It’s who can direct it, structure it and deliver it reliably and responsibly.
The current wave of AI production launches signals momentum, and that’s a positive. But the next phase of the market will be defined less by who can generate something impressive once, and more by who can repeatedly deliver high-quality work under the pressures and constraints of commercial production.
That requires more than tools and more than talent. It requires production thinking redefined for how AI actually behaves. And right now, that’s still the part of the industry catching up.