With our recent additional line of video content, the Behind the Screen blog, I made the attempt to create a more modern and engaging thumbnail. The obvious intent was to draw viewers and drive engagement. In developing it, I did all the usual things – I created mock ups, reviewed similar channels, and watched a dozen or more “how to” videos on what makes a good thumbnail. And in the end, I had all the components – a personal headshot, decent background, strong title, and passable colors. Unfortunately, my composition was way off – too much text, wrong focus, artificial layering, etc. Although I posted it, I then pondered it behind the scenes as “ugly”. In my mind, it was not something that I would have clicked on, so why would I expect others to do so.

Unsatisfied, I tackled the problem a second time, this time looking for more direct help. With the help of some AI prompts, I fed it my initial draft, then asked for suggestions on how to correct the errors. Very quickly, it not only spotted my flaws, but gave me context on how or why I had misinterpreted my research and tutorials. The feedback not only gave me clear suggestions on how to modify things, what to remove, and what to add, but also helped to define some more of the composition. Remove the excess which detracted, and fine-tune what I already had. Then the tool provided a straightforward mockup image of its suggestions, which itself was way better than what I had created.

I could certainly have used what the AI had created – it was functional and professional (and had no typos). But that would only help for this one video, and what I needed was a template for future content. So I went back to my initial design in Photoshop, are started editing again. Very quickly, the 15-20 layers were reduced to 5 or 6 components. I scaled up several pieces, and scaled down a few others. I added some blur to the background, and filled in some of the gaps. I changed the text font, coloring and angles, then did some color corrections to the pieces. And I added a bit of shadow here and there to help the pieces break out of the backgrounds.

A couple of iterations and suggestions later, and I am far happier with the result. The revised image looks far more professional, and gives me a Photoshop format template that I can modify in the future to stay consistent. Yes, the design was assisted with AI, but the final product was all my own effort – and is not a straight copy of the generated version.

Almost certainly my design will continue to evolve – I am not one to just sit with a single idea. As the channel continues to grow, and my experience increases, I will no doubt have a different thought or a different approach. Perhaps I will try a few iterations out, or just re-arrange. But changing something as basic as a thumbnail from “not worth clicking” to “this has possibilities” is a major improvement in my journey.

I know there are many who despise the use of AI for creativity, and many who encourage the use of AI for speed and efficiency. I am more of the middle ground – used responsibly, AI generation can speed production, and provide inspiration and seeds to grow your own creativity. In my case, I could have spent days or weeks trying different things, and maybe at some point I would have come to the same design. But with the aid of my AI coach, it steered me directly to the end goal, while still giving me the understanding to learn from the experience. Neither myself or the AI could have accomplished the goal independently, it required collaboration, which was a win in my mind.