AI video in the UAE: what we have actually used it for, and what we still shoot
Every brand in the UAE has been told AI can make their video for a fraction of the cost. Some of that is true. Here is Rashid, the AI assistant we built for the Government of Ajman in two working days, the three jobs where AI genuinely earns its place, and the ones where we still put a camera in the room.
The question every client is asking, and the honest answer
Somebody has told you that AI can make your video for a fraction of what a shoot costs. The question arrives in almost every conversation now, usually phrased as whether we can just do it with AI. It deserves a straight answer rather than a defensive one, and the straight answer is that it depends entirely on what the film has to do. There are jobs where AI is the right call and we will recommend it ourselves. There are jobs where it would quietly cost you more than shooting, which is the part nobody selling AI video mentions. The difference is not about the technology being good or bad. It is about whether the film needs to show something real. What follows is not a survey of the market or a prediction about where any of this is going. It is what we have actually delivered, where it worked, and where we still put a camera in a room. Most of what is written about this locally is a service page selling AI video, and most of the honest writing is American or Australian and about other markets, so a UAE account of one real job seemed more useful than another overview.
What we built for the Government of Ajman
In August we delivered Rashid, an AI assistant for the Government of Ajman's Department of Legal Affairs. You do not have to take our word for any of this, because the film is on the site: watch Rashid the AI Assistant. It is a short social film, twenty to twenty five seconds, in Arabic, generated entirely with artificial intelligence. We handled the presenter design, styling and background, an AI voice-over recorded to the department's approved script, on-screen subtitles and a light graphic treatment, then sound mix and colour finishing. It was delivered in two working days as a priority service, against the three to five business days a conventional edit takes us, and it carried three rounds of revisions. The client was happy with it and we would take the same approach again on a brief like that one. Two caveats, because one job is one job. The script and brand assets came from the department, which removed a large part of what usually makes a project slow. And the brief suited the method: a short, factual piece of public communication where the point was the information rather than any particular person delivering it. We are describing one engagement that went well, not claiming a general rule. But it is a real answer to the question of whether this works in practice in the UAE, including for a government client, and the answer is that it can.
Where AI earns its place: the presenter film
The clearest case is the one above. A film that needs somebody on screen saying words, where the words matter and the particular person does not. Induction and onboarding pieces. Policy and public information. Internal announcements. Compliance and safety briefings where the content changes every quarter and nobody wants to rebook a presenter and a studio each time. In all of those, the cost of a filmed presenter is real: talent, a day, a room, lighting, sound, and a licence that expires and has to be renegotiated when you want to keep running the film. An AI presenter removes all of that, and for a twenty second internal notice it removes it without costing you anything that brief needed. There is a second advantage that only shows up over time. Because the presenter is generated rather than booked, it can be the same presenter next quarter and the quarter after. That is the thinking behind Rashid: a named assistant the department can keep using, rather than a face hired once. A consistent character is worth something in its own right, and with a filmed presenter you are one resignation or one renegotiated licence away from losing it. The judgement call is whether the person on screen is carrying any of the message. A founder talking about why the company exists is carrying all of it, and an AI version of that is worse than no film. A department explaining a change to a regulation is not, and there AI is simply the efficient way to make it. Ask that question first and the rest of the decision follows. If the film does turn out to need a real person, that is ordinary corporate video work.
Where AI earns its place: inside motion graphics and explainers
The second case is less visible and probably more valuable, because it is not AI making the film. It is AI making elements inside a film we are building anyway. Backgrounds, textures, plates, reference frames, individual assets that would otherwise be illustrated or licensed. On an explainer video or a motion graphics build, a meaningful share of the work is producing assets that nobody will look at twice, and generating some of those rather than drawing them puts the saved hours back into the parts a viewer does notice. The structure, the pacing, the hooks, the timing of a graphic against a line of voice-over. That is where an explainer either works or does not, and no tool decides it. We have said elsewhere that motion graphics is the most important thing after the edit itself, and that holds here. AI changes what the assets cost. It does not change who decides where they go. Nobody should be buying an explainer because it was made with AI. They should be buying it because it explains the thing, and if AI made part of it cheaper to build, that is our problem to manage rather than a feature of the product.
Voice, subtitles and the Arabic version
The third case is the one we would push hardest in this region, because it solves a specifically GCC problem. Nearly every brief here eventually needs a second language version, and historically that meant a second voice-over session, a translator, a separate booking and a separate invoice, which is why so many films quietly ship in English only. AI voice, AI-assisted subtitling and AI-assisted translation have made the second version small enough that it stops being a decision and becomes a default. That is a real commercial change and it is the least discussed of the three. One qualification, and it matters in Arabic more than in English. Generated voice and generated translation still need a human who speaks the language to check them, because register, dialect and the difference between technically correct and actually right are not things a tool is reliable on. The saving is in the production of the version, not in the approval of it. Treat it as a draft that a native speaker signs off, price the review honestly, and it is still dramatically cheaper than the old route. Skip the review and you will publish something slightly wrong in a language your audience speaks better than you do.
What we still shoot, and why
Then there is everything AI does not do, and the line is simpler than the debate suggests. If the film has to show something that exists, we shoot it. Your factory. Your site. Your product in a real hand. Your people, as themselves. A testimonial is worth something precisely because a real customer said it, and an AI version of a testimonial is not a cheaper testimonial, it is a fabrication. The same goes for a founder film, a recruitment film, a facility tour, an event, anything where the proof is that it happened. There is also a quieter reason, which is correction time. When a generated output is nearly right, the work of getting it the rest of the way can exceed the work of having done it properly, and that cost lands on the schedule rather than the quote, which is the worst place for it. That is a judgement from our own projects rather than a measured figure, and we would rather say so than dress it up as a statistic. The practical version: AI is good at producing something that did not happen. It is no use at all for showing something that did. Most brand and corporate briefs need the second one, which is why most of our work is still a crew in a room.
The one thing AI changes about the process: lock the script harder
This is the part clients do not expect, and it is worth knowing before you commission anything. On a filmed shoot, a late script change is expensive but containable. On an AI presenter film the voice-over is generated from the approved script, and the presenter performance is generated against that voice. Change a line afterwards and you are not re-recording a sentence, you are regenerating the piece, which is why we treat a script change after voice-over approval as a revision round rather than a tweak. So the discipline tightens rather than loosens. The script has to be genuinely final, the approver has to be identified before anything is generated, and whoever signs it off needs to have read it out loud. None of that is a drawback, it is just a different shape of project. The pattern underneath it is the same one that governs post production generally: a decision is cheap before the work starts and expensive after it. AI moves that line earlier rather than removing it. If anything, AI rewards clients who were already organised and punishes the ones who were planning to work it out as they went.
How to decide, in two questions
You do not need a view on artificial intelligence to make this decision. You need two answers. First, does the film have to show something that actually exists, or does it have to deliver information. If it is the first, shoot it, and no saving is worth the alternative. If it is the second, AI is worth pricing. Second, is the script truly final and is there one person who can approve it. If yes, an AI piece is fast and genuinely cheap. If no, the correction cycle will eat the saving and you would be better off with a camera and a presenter who can take a direction in the room. That is the whole decision, and we will give you our honest read on which side your brief falls, including when the answer costs us the bigger job. Send us the brief through our contact page and you will get both options priced where both are viable, with the trade-offs written down rather than implied. You can see the range of work either way in our case studies, including Rashid in full, and if the piece you have in mind is animated rather than filmed, that sits with animation production.
