What an AI-specialized production studio is
It's a film production company whose entire process — script, visual design, production and post — is built to work with generative image, video and sound models. That's not the same as a traditional production company adding AI to solve one specific shot, or someone generating one-off pieces with a tool. Artificial intelligence replaces the shoot, but not the creative team: there's still direction, art direction, editing, sound and color.
Put another way: AI is the camera, the locations and the set. Everything else — deciding what the story is, how it looks, which take works and which gets cut — is still craft. An AI-specialized studio should be judged on the second part, not the first.
What AI does well today, and what it still doesn't
What already works very well: scenes that would be extremely expensive or impossible to film, worlds and looks that don't exist, variants of the same piece for different markets without producing again, product photography in any setting, and social content at sustained volume. It also shortens the path from idea to a first watchable version dramatically.
What still takes handcraft, and it's exactly where studios separate from each other: keeping the same face and wardrobe from one shot to the next, reproducing real packaging without the logo or typography warping, micro-acting (a glance, a pause, a hand that grips an object properly) and speech. If a piece needs a specific native accent, for instance, today you have to work with a real voice: no model reliably solves it on its own. A serious studio tells you these things before quoting, not after delivering.
Six criteria to evaluate an AI production studio
1. Finished pieces, not loose shots. Anyone can show five spectacular images. Ask to see finished pieces, with editing, sound and color, and check whether the shots cut together: whether the light, the characters and the product hold from one cut to the next.
2. Character and product continuity. If your piece has a lead character or a real product, ask how they make sure it looks the same in every shot. Working from approved references of characters, locations and packaging before production is a good sign.
3. Human direction and stage-by-stage approvals. A good studio shows you the piece before producing it — script, look, storyboard with reference frames — and asks for your sign-off at every stage. If the process is "send us the brief and we'll send you the video in a week", there's no direction: it's a lottery.
4. Real post-production. A generative model delivers shots, not a spot. Editing, music, sound design, color grading and the final master in every format the media plan needs are half the job. Ask who does them.
5. Rights and commercial use in writing. It has to be clear who owns the piece, for which media and for how long, and how trademarks, recognizable faces and music are handled. If the answer is vague, the risk is yours.
6. Fixed price per deliverable. Billing by the hour in a generative process is a blank check: nobody knows in advance how many attempts a shot will take. The healthy approach is a fixed price per piece, with revision rounds defined from the start.
Questions to ask in the first meeting
Bring these questions to the first call. There are no perfect answers, but the way they respond says a lot: which part of the process does AI do and which part does the team do? How do you keep characters and product consistent? What will I see before you start producing? How many revision rounds does the price include? What exactly do I get: master, cutdowns, formats? And which parts of my idea shouldn't be done with AI?
The last one is the most revealing. A studio that says everything can be done with artificial intelligence is either not being honest, or doesn't know yet.
Red flags
A portfolio made only of stills or three-second shots. Timelines that don't depend on scope ("any spot in 48 hours"). Unlimited revisions with no approval process behind them. Silence about post. Sample pieces featuring celebrity faces or other people's brands, which would be a legal problem in real work. And prices that aren't explained: if they don't tell you what's included, it isn't.
When an AI studio makes sense, and when it doesn't
It makes sense when the idea calls for something hard to film, when you need many versions of the same piece, when the budget can't cover a shoot worthy of the idea, or when there's no time to pre-produce a set. Also as a complement to a shoot: a filmed campaign can be extended with generated pieces that keep the same look.
It doesn't make sense when the value of the piece lies in what you see having actually happened: a real customer's testimonial, coverage of an event, a specific person speaking to camera. That gets filmed. And in between there are plenty of hybrid projects, which are sometimes the best answer. We go deeper in Guide 02: AI advertising or a shoot.
How we work at Dos Lucas
We're an AI-specialized production studio in Buenos Aires, working with brands, agencies and production companies in Argentina and abroad. We make AI advertising, music videos, social content and product photography, always with the same process: previsualization before production, sign-off at every stage, full post and a fixed price per deliverable.
If you're evaluating studios, tell us your idea. Even if we don't end up working together, we'll tell you honestly which part makes sense to do with AI and which doesn't. Everything we do is listed under services.