Neither approach is simply cheaper or better. Traditional filming gives you control over real people, real products and real places. AI generation gives you flexibility to build scenes that would be impractical to shoot, and cheap exploration before you commit. Most good briefs end up as a hybrid, so the useful question is which parts of your commercial belong to which method.
This article compares the two on the factors that actually move budget and risk, then gives a decision table. It deliberately avoids quoting prices: they vary too much by project to be honest in a general guide.
What drives cost in each approach
Instead of asking “which costs less”, ask what each method spends money on.
Traditional filming
- People and time on set. Crew, cast, director, and the days they are booked. Cost scales with shoot days and with the size of the crew.
- Locations and logistics. Permits, travel, set building, props, wardrobe, equipment and insurance.
- Weather and schedule risk. Delays and reshoots are costly because everything is assembled in one place for a short window.
- Post-production. Edit, grade, sound and graphics still apply.
AI generation
- Creative direction and prompt craft. The cost moves from the set to skilled people steering and selecting.
- Volume of generations. Getting one usable shot often takes many attempts. Complex shots take more.
- Compute and tool licences. Ongoing usage costs, and commercial terms that vary by provider.
- Clean-up and finishing. Fixing artefacts, matching colour and rebuilding logos and text as proper graphics.
- Review and legal checking. Extra care on product accuracy, clearance and disclosure.
The takeaway: AI tends to reduce spend on physical logistics and increase the share spent on selection, finishing and checking. It does not erase post-production.
Timelines, control and revisions
Timeline. Filming has fixed milestones: pre-production, shoot, post. AI work is more iterative. Early options arrive quickly, which helps agree a look, but reaching a final cut still needs selection, edit, sound and grade. Do not assume the whole job is faster until the scope is clear.
Control. On set you control framing, performance, lighting and the real product. With generation you steer through prompts, references and selection, which is less precise. Small, specific changes (move the bottle slightly, change the actor’s expression) can be hard to make without regenerating the shot.
Revisions. In filming, late changes may mean reshoots. In AI work, you can regenerate or vary scenes, but each new generation can change things you were happy with. Lock decisions at the storyboard stage and approve in stages. Our guide to storyboarding a commercial covers how to do that.
Realism, talent and real products
Realism. Generated footage can look convincing, but it can also show problems with hands, text, physics and consistency across shots. See the real limitations of AI-generated video for detail.
Talent. Real performers bring authentic performance, and for testimonials or endorsements they are essential. Under the Australian Consumer Law, statements and images must not mislead, and it makes no difference whether the business intended to mislead (ACCC). A synthetic person presented as a genuine customer is a risk to avoid.
Real products. When the audience must see exactly what they would receive, film the product or use real packshots as references. Our piece on AI product videos explains how to keep accuracy.
Disclosure. Platforms have their own rules. YouTube asks for disclosure of realistic synthetic content and exempts clearly non-realistic visuals (YouTube Help). Meta may require a label for photorealistic video created or altered digitally, including with AI (Meta Help). A fully filmed commercial with light AI edits may fall outside these requirements, a fully generated realistic one may not.
Decision table
| Situation | Lean towards | Why |
|---|---|---|
| Exact product appearance is legally or commercially critical | Film (or real packshots composited) | The product must look as it really is |
| Testimonial, founder message or real customer story | Film | Authenticity and consumer-law risk |
| Fantastical, stylised or impossible setting | Generate | Costly or impossible to shoot |
| Many variants of one idea for testing | Generate, then film the winner | Cheap exploration |
| Brand film where performance and craft are the point | Film | Control of cast, light and lens |
| Tight revision cycle on backgrounds or environments | Combine | Film the hero, generate the world |
| Concept pitch before committing budget | Generate as previs | Align stakeholders early |
| Strict regulated claims (health, finance, kids) | Film, with legal review | Higher consequence for error |
Hybrid workflows that work in practice
- Previs, then film. Use generation to build boards and animatics, then shoot the approved plan. It reduces ambiguity on set.
- Film the hero, generate the context. Real product and performer, generated or extended environments.
- Generate the sequence, film the proof. A conceptual opening, followed by real product footage and a real end card.
- Real assets as references. Use real photography and film as inputs so generated shots stay close to the brand.
A simple way to decide
- List every shot in the storyboard.
- Mark each shot as product-critical, performance-critical or neither.
- Film the first two categories unless you have a strong reason not to.
- Consider generation for the rest, and note the disclosure and clearance requirements.
- Get a plan from the studio showing which shots use which method and why.
For short formats, the cinematic 15-second ad guide shows how few shots a tight spot really needs, which can change this calculation.
Questions to ask any production partner
- Which shots in my storyboard would you film, and which would you generate, and why?
- How many revision rounds are planned, and what counts as a revision versus a new generation?
- How will product accuracy be checked against the real item?
- Who owns the final film, the project files and the human-made elements?
- Which tools are used, and what do their commercial terms say about use of outputs?
- What do I receive at the end: master files, cutdowns, licences, a claims file?
A partner who answers these plainly, in writing, is easier to hold to account whichever method you choose.
Where Radiance Impact fits
Radiance Impact is a Brisbane-based creative technology studio, and its Studio produces AI commercials and product films, brand films and trailers, short-form and social video, and post-production. If you are unsure which method suits your brief, contact us to talk through the options, or see the Studio page.
