Can I build a commercial myself with Kling, Veo or Sora?

Technically yes, the tools are freely accessible, and anyone who reads up will generate their first clip tonight. But the effort lies less in the generating than in concept, consistency across several shots, editing, sound and rights clearance. On top of that comes the reject rate: for our Manner spec, prompting alone took about six hours for 40 seconds of film, and the viral Kalshi spot at the NBA Finals needed 300 to 400 generations for 15 usable clips. So for social experiments, internal videos or just trying things out, doing it yourself is often enough. But as soon as media budget sits behind the video, several spots have to fit together consistently, or a deadline forgives no learning curve, experience pays off.

What are the most common DIY mistakes?

The three classics: people generate straight away without a script, the clips don't fit together visually, and sound is an afterthought. All three are plannable craft questions, not tool questions. Without a concept you get pretty individual images but no story; without a system of reference frames, characters and look drift from shot to shot; and without sound design even good footage feels like a tech demo. On top of that comes the underestimated reject rate, which costs beginners time and nerves. The good news: because these are craft problems and not tool problems, they can be avoided with planning, no matter which generator is currently the best.

Is a hybrid of in-house team and pro worth it?

Often yes, and in practice it's one of the most sensible routes. Concept, visual world and the decisive key shots come from the pro, while the ongoing cutdowns and social variants are made in-house by your team. That gives the brand consistency and a professional foundation without every small thing having to be outsourced. It keeps quality high where it's visible and ongoing costs low. Especially for companies that need content regularly, this middle path is often cheaper than either extreme: completely self-made and therefore inconsistent, or everything external and therefore expensive. What fits in your case depends on how much experience and time you have in-house.

Guide · Decision

AI video: do it yourself or hire a pro?

An honest decision guide from someone who makes a living being hired, and still tells you first when you are better off doing it yourself.

Last updated: July 2026

First, so the rest stays credible: yes, sometimes DIY is enough. The tools are more accessible than ever, and anyone reading this can generate their first AI video tonight. The honest question is not whether you can, but when your time is well spent and when it is not.

When doing it yourself is perfectly fine

For social content with a short half-life, internal videos, moodboards and pitch visuals, or simply to learn: no agency and no filmmaker needed. A tool subscription and a few evenings of practice will do. It is valuable groundwork too: whoever has generated themselves will brief far more precisely later.

What DIY tends to underestimate

The rejection rate

The viral Kalshi NBA Finals spot took roughly 300 to 400 generations for 15 usable clips. Our own Manner spec took six hours of prompting and generation alone for 40 seconds of film. The rate improves with experience, but it never disappears.

Consistency

A pretty single clip comes quickly. A character that looks the same in shot 7 as in shot 1, a product that matches yours exactly, one continuous lighting mood: that is systems work with reference frames and planning, not prompt luck.

Rights and commercial use

May the tool’s output be used commercially, and to what extent? How is the music licensed once the spot runs as an ad? Irrelevant for private experiments, essential for brand communication.

Sound and editing

The difference between “AI clips in a row” and a commercial is made in the edit, the sound design and the music. These are classic craft disciplines no video generator ships with.

When you should hire a pro

As soon as your brand is visible, media budget sits behind the video, several spots must fit together consistently, or a deadline exists that forgives no learning curve. Rule of thumb: the higher the visibility and the more recognisability matters, the more experience from professional AI film production pays off.

There is also a middle path that works well in practice: concept, visual world and key shots from the pro, cutdowns and social variants made in-house by your team. Quality stays high where it is visible, ongoing production stays affordable.

What hiring costs depends on a few clear factors: how much does an AI commercial cost?

FAQ

Frequently asked questions

Can I build a commercial myself with Kling, Veo or Sora?: Technically yes, the tools are freely accessible, and anyone who reads up will generate their first clip tonight. But the effort lies less in the generating than in concept, consistency across several shots, editing, sound and rights clearance. On top of that comes the reject rate: for our Manner spec, prompting alone took about six hours for 40 seconds of film, and the viral Kalshi spot at the NBA Finals needed 300 to 400 generations for 15 usable clips. So for social experiments, internal videos or just trying things out, doing it yourself is often enough. But as soon as media budget sits behind the video, several spots have to fit together consistently, or a deadline forgives no learning curve, experience pays off.
What are the most common DIY mistakes?: The three classics: people generate straight away without a script, the clips don't fit together visually, and sound is an afterthought. All three are plannable craft questions, not tool questions. Without a concept you get pretty individual images but no story; without a system of reference frames, characters and look drift from shot to shot; and without sound design even good footage feels like a tech demo. On top of that comes the underestimated reject rate, which costs beginners time and nerves. The good news: because these are craft problems and not tool problems, they can be avoided with planning, no matter which generator is currently the best.
Is a hybrid of in-house team and pro worth it?: Often yes, and in practice it's one of the most sensible routes. Concept, visual world and the decisive key shots come from the pro, while the ongoing cutdowns and social variants are made in-house by your team. That gives the brand consistency and a professional foundation without every small thing having to be outsourced. It keeps quality high where it's visible and ongoing costs low. Especially for companies that need content regularly, this middle path is often cheaper than either extreme: completely self-made and therefore inconsistent, or everything external and therefore expensive. What fits in your case depends on how much experience and time you have in-house.

Your Project

Rather talk directly than keep reading?

Start a project