Using AI Inside a Brand Film Without Letting It Touch the Logo

AI image and video tools are good at atmosphere. They can produce a warm, snowy, festive scene quickly, with depth and detail that would take a long time to build by hand. What they are not good at is a logo, because a brand mark has to be exact, and a generator only approximates.

So our rule is simple. AI can make the world, but it must never draw the brand. In our festive animation of the Marriott M, the scene around the logo was generated, and the logo itself was animated from the supplied artwork.

The first step is to agree where the boundary sits. We separate the film into layers: the generated background, the supplied logo artwork and any text. The logo layer is always built from the client's official files, so its shape, colour and spacing are correct to the pixel.

The second step is to make the generated scene respect the logo. We leave clear space for the mark, check that it has enough contrast against the scene, and avoid generated elements that look like part of the logo or imitate it. An AI-made snowflake that happens to resemble the brand shape is the sort of accident we look for.

Third, we choose the generated material with care. Odd details such as warped hands, nonsense text or wobbling edges are easy to miss on one pass, so we review every second of the output at full size before it goes into the edit. If a frame fails, we regenerate or cut it.

Finally, we tell the client what we have used. Brands increasingly want to know where AI is involved, and being open about it builds trust. We describe which parts were generated, which were built by hand, and which came from their own assets.

Used this way, the technology saves time on the parts that benefit from it and leaves the brand's identity untouched. It is a sensible split for any seasonal film. You can see our Marriott work on our case study page.