Most of the conversations we have with brand and agency teams begin the same way. They have heard about AI video production, they have seen reels, and they are trying to work out where it fits in their roster. The honest answer is that it fits in some briefs brilliantly, in others not at all, and in most briefs it sits alongside traditional production rather than replacing it.
This guide is the framework we use internally to triage briefs. It is opinionated, designed to give a clear recommendation rather than a maybe, and grounded in the work we actually take on for global brands.
The three-question test
Before any other consideration we ask three questions about the brief. The answers usually point to one of three buckets.
- Is performance the idea? If the joke, the tear or the moment of recognition is the ad, you need a real human in front of a real camera.
- Is a recognisable real person or place essential? Real celebrities, athletes, presenters, your actual factory, the actual high street. Hard to fake well and risky to fake at all.
- Is the schedule a hard constraint? Testing date, air date, pitch deadline. If the date will not move, schedule weighs everything else.
Two yes answers usually means traditional production. Two no answers usually means AI. A mix usually means a hybrid pipeline where AI does the work it is good at and traditional shoots the work it is not.
When AI video production is the right call
Research animatics
If you are testing creative routes before committing production budget, AI animatics deliver in 5 to 10 working days at a fraction of traditional cost. This is the use case where AI has the largest and most defensible advantage.
Multiple route testing
When the brief is to explore three or four directions rather than execute one, AI lets you produce all of them for the budget of a single traditional shoot.
Multi-market versioning
Producing twelve culturally adapted versions traditionally is a procurement nightmare. With AI it is a one-week production run after the master is locked.
Pitch and pre-vis
If you are pitching an idea and need to show what it looks like rather than describe it, an AI film made in a week is a better selling tool than a deck.
Impossible or expensive worlds
Scripts that call for environments you cannot get permits for, weather you do not have time for, or eras you cannot stage are cheaper and faster to generate.
Social and digital at volume
Variant heavy social and digital content scales naturally with AI in a way traditional shoots do not.
When traditional production is the right call
Performance-led hero ads
When the script depends on a specific actor's delivery, the live action craft is irreplaceable. AI casting can carry mood and presence, not a comic beat or a tear.
Real celebrities, athletes and ambassadors
Likeness rights and audience trust both favour shooting the real person. AI talent is for invented characters, not real ones.
Hero broadcast spots scrutinised frame by frame
For your once-a-year flagship, the marginal craft of a top traditional production is worth the cost. Use AI in pre-vis and pre-test instead.
Intricate product hero shots
Product packaging, label legibility, intricate hand-product interaction. Easier to shoot than to generate cleanly.
When a hybrid pipeline is the right call
Most flagship campaigns we work on now end up here. The pattern is consistent: AI animatics for route testing, traditional shoot for the hero, AI for variants. Each step plays to the strength of the production approach beneath it. The cost is materially lower than a fully traditional pipeline, the calendar is materially shorter, and the creative ceiling on the hero spot itself is uncompromised.
We cover this sequence in more depth in the script-to-screen workflow. The summary is that production used to be one decision and is now three smaller ones, each suited to whichever pipeline fits the deliverable.
Budget and schedule reference points
As rough guideposts, a research animatic is a 5 to 10 working day project at £4k to £15k per route. A single-market AI hero film is a 2 to 4 week project at £15k to £45k. A multi-market AI campaign with versioning is a 4 to 8 week project at £40k to £120k+. A traditional broadcast hero starts at £150k for a single market and rises sharply from there. Our full pricing page goes into the specifics.
If you compare those numbers against a brief, the right call usually becomes obvious. Hero broadcast at flagship budget, traditional. Multi-route testing or multi-market variants at quarterly budget, AI. Everything else, a conversation worth having before quoting.
Red flags that suggest you should not commission AI
- The brief is built around a specific person whose face and voice are the campaign.
- The film opens or closes on a tight close-up of a hand interacting with the product in a complex way.
- Regulators require a documented real-person testimonial.
- The brand book mandates only photographed talent and product.
- The budget is below £4,000 and the scope is genuinely simple.
How we triage a brief
When a brief lands we run the three-question test, look at the deliverables list, and then ask one more question: what happens if this fails research or fails to clear. If the answer is rework, AI helps because the cycle time is short. If the answer is reshoot, traditional is fine because you would not be re-running the AI pipeline either.
If you would like us to run the same triage on a brief you are working on, send it over. We will tell you which pipeline we would recommend and why, even if it is not the one we would quote on.
Related reading
Want a take on a specific brief?
Tell us what you are producing and we will tell you which pipeline we would recommend.
About this article
Written by James Finlay, Creative Director at Myth Labs. Reviewed for accuracy by Izzy Hill, Head of Client Success. Based on our production experience and industry research.