In the era of advanced image generation and Claude Code powered workflows, it’s undoubtedly hard to resist the urge to include ourselves in our own design process.

Niko, our internal design lead at Diligent Studios, has done his fair share of tool exploration when it comes to using AI in ad design and these are his thoughts:

✦ I never try to one-shot the entire creative. The idea that a single prompt and output will magically turn the image in your head into a finished ad rarely works. Instead, I use AI to generate specific parts of the creative: the focal image, background, or small details. Then I combine those pieces into the final composition in tools like Affinity or Figma (yes, these are still needed). Think of it more like a collage workflow where AI helps create individual layers.

✦ Different models produce different results. When generating certain parts of a creative, it helps to use workflow-based tools like Weavy/Freepik to test multiple models at once. Some models work better depending on the style or use case. Experiment and explore instead of settling for the first usable output.

✦ I always create lo-fi sketches before moving into hi-fi design. That includes the copy, rough grayscale layouts, and composition. But nothing is set in stone. Want to see the focal image from a different angle? Ideate with AI. Curious how it looks with more dramatic lighting? Explore it. The goal is to experiment more, not less, with the tools available to us. Oh, and did I mention not to use that as the final output?

✦ Last but not least, you don’t need AI at all to design great ads. Seeing endless AI workflows, prompts, recipes, and prompts floating around shouldn’t stop you from sitting down with pen and paper (or firing up Illustrator/Affinity/Figma) and doing it the old-fashioned way. That’s where creativity actually lives and breathes. It’s in the hours spent ideating, doubting yourself, and searching for new ways to communicate an idea through your own point of view.