What an AI commercial costs, and what moves the budget
If you hold the budget, someone has probably told you AI makes the job cheaper. This explains what an AI production still pays for and the six questions that set the size of an estimate, so you can read one and see what is moving it.
An AI commercial is not automatically cheaper than a shoot. AI can reduce some of the physical costs, such as locations, sets and crowds. Planning, direction, control, revisions, finishing and delivery all remain.
01 · The Drivers
Six things that decide an AI production budget
The price of a generation, the length of the prompt and the name of the model are small parts of the budget. The larger part is keeping approved decisions fixed while everything around them changes. These six questions set the size of that work, and I ask all of them before I estimate anything.
1. How many shots must match
One striking image is the inexpensive part. A commercial needs shots that agree with one another: the same character, the same light, the same world from another angle. Each shot that has to agree with the others adds control work on top of generation. That work is how consistency is held across a sequence.
2. Which elements must stay exact
Talent, product and logo cannot drift. The product has to remain correct and the brand recognisable in every frame, so those elements are separated from the parts where generation is allowed to explore. Objects and characters that must persist are often built in 3D. The more of the frame that has to stay exact, the more of the budget goes on control.
3. How much live action has to be integrated
If real actors or filmed plates are part of the job, the generated material has to match them in lens perspective, light, scale and contact shadows. That is compositing and VFX work, and it is planned before the shoot. It is part of the VFX work an AI shot still needs.
4. How many rounds, and how many approvers
The first review responds to the general idea. Later rounds get specific, and brand guidelines and legal come back into the discussion. Each approved element becomes something that can no longer move, so every later version has to respect a longer history of decisions. The number of rounds, and the number of people who can ask for changes, has to be agreed and scoped at the start. This is why "change one thing" is the hard request.
5. How many deliverables and versions
One film in one format is a different job from a campaign with several delivery formats, variants or languages. A workflow that has been built and tested once can be run again for the next variant, which is where versions become affordable. Each one still needs finishing and checking, so they belong in the estimate from the beginning.
6. Whether cloud models are allowed
Some briefs allow AI services that run on someone else's servers. Others require your material to stay on controlled hardware, which means models that run locally. That changes the tools, the setup and the cost of each iteration, so it needs to be known before the estimate. More on that in how the model for each shot is chosen.
02 · The Brief
The budget edits the idea before the shoot
Before a commercial is made, the budget has already started editing the idea.
The crowd gets smaller. The location gets closer. The set becomes a room that already exists. The ambitious scene becomes a suggestion of that scene. The campaign becomes one piece of work instead of a larger set of assets that could meet people in different places and in different ways.
None of this is a failure of imagination. It is how production works. People need to be cast, moved, fed, dressed, lit, filmed, insured and sent home again. Locations have to be found, hired, prepared and restored, and every additional day adds to all of it.
AI changes some of those decisions. It can remove the right physical constraints from a brief and return some room to the idea. But it also creates a temptation to ask the wrong question.
03 · The Cheap Version
Cheaper is not always a saving
The first question around AI is usually, “Can we make this cheaper?”
Sometimes it can, and that is the least interesting thing about it. The low cost of a single generated image is the visible part of the calculation. It is also why so much work now looks convincing for a few seconds and becomes difficult to watch after that.
A campaign is judged by what happens after people see it: whether the product looks right, whether the message arrives, and whether it gives someone a reason to trust the brand or buy from it.
If cutting the production cost cuts the credibility of the work, the attention it earns, or the confidence people have in the company behind it, then the saving appears on the estimate and disappears everywhere else.
04 · The Same Budget
What the investment can now buy
The more useful question is what the same investment can now do that it could not do before.
A conventional production spends a large part of its budget making a physical day possible: travel, access, locations, sets, equipment, and getting every element into one place at one time.
Some of those costs are essential. Some are the cost of the old route.
When AI removes the right constraints, more of the investment can go into the work itself: the direction, the quality of the frame, the worlds the story can enter, the finishing, and the campaign material that gives the message more ways to reach people.
The aim is to give the original idea more of a chance to become what it was meant to be before the budget began reducing it. That is different from making more output for the sake of it.
05 · Same Preparation
The work moves. It does not disappear.
There is a version of AI production that stops at the first attractive result. It treats the generated frame as the finished work and assumes the rest will follow. It rarely does.
A proper production still needs an idea, direction, taste, references, decisions about what must remain fixed, and a route for the revisions that are always going to arrive.
AI changes where preparation happens. It does not make preparation less important.
That is why a cheap AI process can become expensive later. A project can generate endless options and still have no reliable way to carry an approved decision into the next image, the next shot, or the next client note. The work looks inexpensive because each generation is inexpensive. The rebuilds are not.
06 · The Difference
A shortcut or a production system
AI can create worlds, assets, tests, variations, and scale that a conventional budget may not have reached. It becomes useful in a serious production when those things can be directed, controlled, revised, and finished properly.
That is where AI is combined with the rest of the production process: 3D, visual effects, compositing, automation, and the ordinary judgment required to decide what the audience needs to see and what the client needs to approve.
The crowd on Gigaclear is one example. A physical version would have consumed a substantial production day. It was built as a controllable system instead, and the full account is in the crowd that did not need a shoot day.
What matters there is the decision behind the crowd: identify where the old route is expensive, understand what the finished work requires, and build a better route before the budget is committed.
07 · The Return
A bigger chance to achieve the goal
I judge an AI production by what the investment returns, and the size of the estimate comes second.
With the same investment, a campaign can have more ambition and more visual control, and the idea has more chance of producing the result the campaign is measured on.
That does not guarantee a result. The work still has to hold attention, communicate clearly, and be credible enough for the brand to stand behind it.
But a serious production has a better chance of doing those things than a cheap version made only because it was cheap. The budget does not become less serious because AI is involved, and neither does the preparation. What changes is the size of the goal the same investment can pursue.
If you want the six drivers worked through for a specific script before any budget is committed, send me the script. The first read is free, and you get a written reply within a day.
Frequently Asked Questions
Does AI production always cost less?
No. It can remove expensive physical constraints, but a serious production still needs planning, direction, control, revisions, finishing, and delivery.
What decides the budget of an AI commercial?
Six things: how many shots must match, which elements must stay exact (talent, product, logo), how much live action has to be integrated, how many revision rounds and approvers there are, how many deliverables and versions are needed, and whether cloud models are allowed or the work must stay on local hardware.
Why can cheap AI be a false economy?
Because lower production cost can also mean lower credibility, weaker audience response, and a campaign that does less of the job it was made to do.
What does "Bigger vision, same budget" mean?
It means using the same investment to give the original idea more scale, more visual range, and a better chance of reaching the campaign objective.
What does preparation mean in AI production?
It means deciding what needs control before generation begins: the message, brand, product, visual language, approval path, and revision path.
Send the script and the budget you are working to. Within a day you get a written reply: where AI can reduce the cost of the physical day and where that money would do more. This first read is free.
Send the brief → See how the Gigaclear commercial was made →