Date: Early 2023

One of my biggest passions is to constantly look at new ways of doing something with the purpose of improving efficiency in my workflow as well as staying creative throughout. I maintain a growth mindset by experimenting with new products, software, and methods. I have grown quite curious about AI + Product Design ever since the introduction of Chat-GPT. Since then I have tried and tested several AI tools in depth to grow my understanding

Such as the following:

Chat-GPT (Text)

Bard (Text)

Midjourney (Imagery)

Luma AI (Video Capture)

Runway (Imagery)

AI Test Kitchen (Music)

Gamma (Presentation)

Tome (Presentation)

WonderDynamics (Film, VFX)

While there are a variety of ways to go about that I choose to build this app using a new software called 'imagica' - a no-code AI App Builder.

I am currently consulting for impacteers.club, an indian startup that is building a new career guidance platform. While I am focused on improving the product itself, I was asked to provide some guidance to the video content creation team (due to my background in film school and as content creator).

I thought this would be the perfect opportunity to create a tool that can help the team improve their social media presence in a specific area.

The project was completed within a 1-day time frame.

The team often struggled to select the ideal thumbnail for their content possibly resulting in a low view count. Furthermore this process is typically time-consuming, requires a certain level of graphic design skill, and there is no guarantee that the chosen image will resonate with the target audience.

My objective was to develop a solution that simplifies the thumbnail selection process, while simultaneously increasing video engagement through more effective thumbnail choices.

This tool should be extremely easy to use and under no circumstance slow down the teams content creation process. Hence I wanted to keep it simple. What does it need to accomplish?

Provide a description of a desirable thumbnail based on a given topic

Educate user on how to achieve such a thumbnail

The user will only need to input the topic of the video. The AI will take the information and generate the expected info based on a trained set of information.

Through prompt engineering and a node based editor I built the app.

Let's test it out.. The team was in the midst of creating content on helping students prepare for job interviews.

We simply enter the topic of the video:

I sent out V1 to the marketing team in hopes it would improve their workflow. The feedback was positive but I notice that they struggled with visualizing the given thumbnail.

Back to the drawing board...

What does it need to accomplish?

Provide a description of a desirable thumbnail based on a given topic

Educate user on how to achieve such a thumbnail

As image-generation models struggle with text I have chosen to only generate a blank image that can be used as background inspiration.

The tool has been happily adopted and is currently being used by the social media team .

I also posted the tool in a design community and received positive feedback

Imagica is currently in an early beta and as such has quite limited functionality. I'm looking forward to improve my tool as time goes on.

You can test out the tool yourself: Find the Perfect Thumbnail