We recently put a range of AI tools to the test with a simple but meaningful challenge: create real advertisements that could run on a publisher’s website or appear in print.
The lineup included Adobe Firefly, ChatGPT, Copilot, Canva, Gemini, Reeve and others. The results were mixed and, more importantly, they were revealing. (The link to the results can be found at the bottom of this article.)
As AI continues to evolve, its role in creative production is becoming harder to ignore. For publishers and advertisers alike, the promise is clear: faster turnaround, scalable creative and reduced production costs.
But in practice, quality varies widely.
Some outputs were polished and nearly ready for use. Others missed the mark entirely. The difference often came down to one critical factor: how the AI was prompted.
One of the most important lessons from this exercise is that prompting isn’t just a step in the process. It IS the process.
In fact, one of the most effective workflows we identified flips the traditional approach:
First, ask the AI to generate a detailed prompt based on the desired ad specifications (size, tone, format, audience). Then, use that AI-generated prompt to produce the final creative.
This two-step method consistently improved output quality, often resulting in ads that were usable — or very close to it.
A common question is: Which AI tool should you use?
The honest answer is that it depends.
Different models excel in different areas. Some are better at layout and visual composition. Others perform well with copywriting or concept generation. But no single platform consistently delivers the best results across every scenario.
For publishers trying to generate high-quality ad creatives for both print and digital, this creates a challenge, and also an opportunity.
Rather than committing to a single AI model, a more effective strategy is to remain flexible:
This approach not only improves results but also provides greater control over the creative process.
AI-generated advertising is no longer theoretical. It’s already being integrated into real production workflows. As tools improve and prompting techniques mature, the gap between “AI-generated” and “production-ready” will continue to shrink.
For publishers, the focus must shift from whether to use AI to how to use it effectively.
And as this space evolves, one thing is clear: the future of ad creation won’t be powered by a single tool, but by a smart combination of them.
You can check out the results of our tests and the prompts we used at this link.
Kurt Jackson is a newspaper industry veteran of 40+ years and is the owner and managing member of Software Consulting Services (SCS). SCS offers an extensive line of publishing-related applications, including Layout-8000. SCS’s mission-critical software is used every day by more than 500 sites in 10 countries producing over 16,000 publications in five languages. SCS also resells award-winning digital asset management software from FotoWare based in Oslo, Norway. Learn more at www.newspapersystems.com.
