The AI hype seems to be continuing. We are dabbling with it in several areas of our business and will let you know if we find any interesting use cases. Lately, we used Copy.ai to conduct a test for generating content on advertising texts in Facebook Ads.

How did we do it?

Basic data:

  • Ad format: animation
  • Campaign objective: traffic
  • Duration: 30 days
  • Allocated total budget: 200 euro
  • Number of ad sets: 3

The main goal of creating the social media content was to promote a consulting service and prompt users to visit their website. We planned to measure the effectiveness of the campaign based on the CTR indicator.

In each group of ads, a creative was added with copy written by a human and  copy generated by AI to test how each description would affect the overall perception of the ad. The animation, placements, and CTA in both versions were the same.

The results varied by ad group, but in ⅔ of the cases, the creatives with human-generated text had a quite significantly better CTR.

However, we suggest treating this test more as a curiosity than inspiration for you to try it. That said, we will most likely use AI prompts again.

However, a few conclusions can be drawn from our micro test:

  • AI can be a great help in the marketer’s work, but it will not do the job on its own, at least not yet. 
  • Not all the generated texts were coherent and/or linguistically correct. The text selected above was the only one suitable for using in advertising.
  • The tool chose the tone of the communication appropriately – it was not too formal or too friendly and suited the format of the Facebook post.
  • It reduces the time needed for research and speeds up the process of creating several copy versions.
  • If we make minor changes to the content, the user reading the post will not be able to distinguish whether it was written by an AI or by a human.
  • AI is great for ideation but it will not add character to the content, it will not use jokes or irony that so often make messages, not only advertising ones, more attractive.
  • What AI will not leave room for mediocrity. With the development of AI, there are greater demands on employees of the creative industry, not only in terms of the development time for materials but also in the quality of their implementation.
  • AI vs. company confidential information – we don’t have specific information about what’s going on with the data we process. It is better not to include sensitive company data in the company’s tools, at least at the moment.

There are more and more tools that use AI. Here is a short list of tools that may be useful in marketing departments:

  • ChatGPT is probably the loudest, and no wonder, since it is a real threat to Google.
  • Runway helps you with video editing.
  • Galileo independently designs user interfaces based on a short description.
  • Deepl translates over 31 languages ​​with more accuracy than Google Translator.
  • Rytr.me generates many categories of content from shorter forms like CTAs, posts, and product descriptions to e-mails and blog entries.
  • Fireflies helps you create meeting notes.
  • Tome.app creates visually appealing presentations. Will Canva get left behind?
  • Futurepedia.io is a constantly updated search engine for AI tools.

Do you already use AI tools in your daily work? Already have your favorites?