Businesses across the globe are buckling to the pressures of incorporating AI into their wheelhouse for a chance to be part of the billion-dollar industry revolution. Here at 106 Comms, we’re embracing the opportunity to cut our teeth on the future and find out the best way to utilise this technology. Here we detail some of our trials and results from experimenting with some of the big hitters leading the way in AI content creation.
To start with, we kept with the new classic of ChatGPT. Following the advice from a few different seminars and my own reading, I wanted to test how much it helped with moving passed the blank page. This began by inputting a version of a speculative brief into the software and asked it to produce some straplines, taking out any brand names so as to keep the pull as original as possible. With a few different inputs, I found that it could be clinical and a little overly wordy but nothing that a few extra inputs couldn’t fix. It could make it more concise as well as fold in key phrasing. In all honesty however, the auto-generated concepts were a starting block for full copy but it did throw up unique phrasing that could be built upon into some tangible concepts.
In another vein, I used ChatGPT as an interpreter of existing copy to find a more unbiased viewpoint than my own. I took early drafts of work and asked ChatGPT to give its interpretation of the copy with no other directions added. I was hoping to use it as a means of ensuring top-down messaging. The analysis could sometimes be quite one note but helped to ensure core focuses and made me more confident in bringing deliverables to future clients.
A benefit of ChatGPT is much more outside of copywriting itself and can actually used as a display tool. This was one of the more successful trials as on a personal note, comprehending what data points mean and folding them into creative was something that requires more effort for me. Because we concern ourselves with employer branding, I decided to reacquaint myself with the latest statistics on active job seeking and company investment into employee wellbeing. I input this into the system and asked ChatGPT to present it in tables and provide broader insight to find aspects to pull from. Not only did it make the information much more accessible but also made some early calculations to understand next steps with the data presented. More specifically, it made some rough calculations on the savings that companies can make by investing in employer wellbeing.
In terms of more wide-ranging content, I turned to Runway to have a crack at generating photo to video content. For the full experience, this required a month-month subscription but there’s also enough to experiment with a few different videos on the free version if you’re not ready to fully commit. It runs on a point-based system that get used with each prompt, which is oddly helpful for making sure your prompts have maximum impact.
I learned quickly to build upon incremental gains as some of the more ambitious prompts tended to overwhelm the software. I took a variety of different photos of me and decided to try and put them together in a cohesive narrative. One of the tests I ran was a photo of me on stage, holding a microphone, which I wanted to turn into a video of me telling a joke and laughing.
This took a lot more returns to the drawing boards than expected as I continued to hone the specificity of the prompts. Some of the leaps that Runway’s software were bold but very much not the ideal. I could see aspects in the finished product when they tried to match my facial features with their existing database, even giving me a very generous amount of stubble that I do not have. Luckily, you also have the opportunity to return to former iterations and make the adjustments necessary to redirect for better results. It’s also worth remembering how hands off my inputs were as I only used a maximum of three photos to complete the full project.
By and large, using this software did require patience. It’s very much based in how you interpret your results and the design decisions you make. It’s not at a point where you can request exact results while you slink away to make a coffee. However, the inspiration points it pulled out in both Runway and ChatGPT worked effectively as a jumping off point and consistently refocussed creative efforts when certain aspects got lost in the weeds.
More than this, it’s been made incredibly user friendly and the accessibility proved to make the whole enterprise a lot less intimidating than first expected. Giving it the time of the day to parse the results is well worth the effort and introduces an extra angle on any brief that sparks the final ideation you need. In the hands of deft and curious creatives, it can be used consistently to break idea ruts and reframe your perspective.