Let us take a preparation from a TV magazine as a finger exercise.
It can take some time before such an ad is created, because headlines and image motifs have to be found that fit the target group of the preparation. Artificial intelligence can - already today - support these steps by developing suggestions from which the human user:inside can then help themselves.
Creating a good headline is often an extensive process that requires a lot of trial and experimentation. While it's unlikely that an AI will formulate the final headline, you can expect your agency to provide a wider range of variations at the beginning of the project. Here's an example:
Prompt for ChatGPT:
Write me 5 different variations of the headline of an ad for a homeopathic preparation that works well against pain in joints, especially in wear and tear diseases. It specifically combats osteoarthritis, is well tolerated and can be dosed individually because it is taken in drop form. Write the results in a table with two additional variants "funny" and "serious".
The headlines are not yet final, but can be a good starting point for discussions about which direction to go. So don't hesitate to be demanding of your agency when it comes to getting a variety of proposals in these early stages.
Style finding for a product image
In the style discovery phase, your agency can create different variations of a product image. A simple example of this comes directly from Midjourney:
Again, of course, the images are not final yet. However, it is quite easy to create 20 or more images, try out different styles, and then discuss which direction the design should take based on a variety of suggestions.
Possible segments of a target group
Why don't you ask ChatGPT which target groups are suitable for your product? For our example product, ChatGPT suggests the following target group segments:
Are these definitive segments? No. But again, such an answer can serve as a starting point for AI to select target groups and validate sizes. More examples could be run for market research, long copy, or ad variations. The pattern of processing and the basic idea of getting loaded with basic info and ideas by the AI remain the same.
Things you can not expect
A few projects may become slightly cheaper in price by using an AI, but our experience so far shows that this is rather rare. It still takes time and resources to evaluate the results of the AI and to generate high-quality results from it.
For you as a customer, the demand for the wealth of variants and sometimes even the number of elaborated proposals to be presented in the team increases. However, this can actually lead to higher costs than lowering the price.
Moreover, you should not expect finished results from an AI - at least not yet. At the moment, these results are all good interim results that can be used as a basis for further work. For us in Rx marketing, these results are still a long way from meeting the requirements of something good or finished.
It pays to demand richness of variation from your agency at the beginning of a phase to have a broader basis for decisions. Don't hesitate to communicate these expectations clearly. You may even enjoy trying out different options with agencies and comparing them. Remember that the results of AI-based tools are not yet finished outputs and need further work to meet the high expectations in Rx marketing.