Campaign production with AI is the topic many marketing departments talk about and few have properly calculated. The tools get better every month. The question is no longer whether they help in production, but where in the chain they really save time and money.
I look at it from the perspective of someone who has built many campaigns. At a fashion house I was responsible for 40 photoshoots, 6 TV commercials and 12 drive-to-store campaigns. Today I lead a team of 22 people across markets in EMEA. If I planned these productions again today, AI would be part of some steps and deliberately left out of others.
- In campaign production, AI saves time mainly on variants, formats and adaptations.
- The idea, casting and approval remain human work.
- Without clear brand rules, AI produces faster, but not better.
- The bottleneck moves from production to approval.
Contents
Where campaign production with AI saves time
A campaign rarely consists of one visual. One photoshoot produces dozens of formats: portrait, landscape, square, with and without text, for shop windows, newsletter, social and online shop. These adaptations used to be manual work in production and often took more time than the shoot itself. This is exactly where campaign production with AI is strong today.
The same applies to language versions, text lengths and first drafts of storyboards. A team that hands this work over to tools gains time for the questions that make a campaign good. McKinsey estimates that about three quarters of the value of generative AI lies in four areas, and marketing and sales are one of them (see here: McKinsey, The economic potential of generative AI).
For campaign production with AI the order matters: describe the process first, then choose the tool. Teams that start with the tool often automate exactly the steps that were never the bottleneck.
An example from practice: a drive-to-store campaign needs posters, windows, ads and in-store content alongside the key visual. Preparing these adaptations with AI and having people approve them saves days in every campaign.
What stays human in campaign production
The idea behind a campaign does not come from variants. It comes from a point of view on the brand and a sense of what is happening in culture at that moment. That is the work of people who know the brand and its audience.
Casting, location and the mood on set cannot be replaced either when a brand lives on real people and real products. In fashion especially, the product has to be right: colour, material, fit. A generated image that shows a garment wrongly later costs returns and trust. More on this in the article AI in image production.
Brand rules as the foundation
AI produces as well as the guidance it receives. A brand whose rules only live in the heads of a few creatives becomes random faster with AI. That is why clear rules for imagery, tone of voice, typography and taboos belong at the start of every AI production.
These rules have to be written so that both a person and a machine understand them. I describe what that can look like in the article Brand guidelines for AI.
The new bottleneck: approval
When variants cost almost nothing, more of them get made. That sounds good, but it moves the bottleneck. Suddenly there are not ten but a hundred visuals waiting for sign-off. Without clear criteria and clear responsibility, approval becomes the slowest step in the whole chain.
That is why I explicitly plan time and people for review in every campaign production with AI: is the product right, is the brand right, are the legal points right? Thinking about this from the start is what really saves time.
My conclusion: campaign production with AI pays off where a lot is repeated, and it fails where the brand is decided. Good teams use the time they gain for the idea and for approval.