At this point, the question isn’t whether artificial intelligence should be used for creativity – it’s already happening. The real questions are, “how do we use it?” and “how do we do it right?”
There are genuine and valid concerns around the use of AI in creativity, and all of them need to be considered. Equally, we can’t ignore that AI, especially generative AI, is now part of the creative landscape. We need to be on top of it, using it in the right ways and the right places, so that we’re not left behind.
Let’s look at CRM specifically – partly because it’s what we do at Armadillo, and partly because the topic of generative AI is way too big to cover in a blog post without narrowing our focus to one area.
In CRM, we’re talking about a channel that has conversations directly with consumers and is designed to build and strengthen customer relationships. How can artificial intelligence build a real human connection? Ultimately, it can’t. We’re talking to humans, so we still need human creatives injecting some fun and passion into our communications.
So, where do we find room for AI?
Right now, AI can’t innovate – it’s not true ‘intelligence’, it’s machine learning. ChatGPT can’t come up with new ideas, it can only reframe and recombine content that’s already been created. And, while it can do that in interesting ways, it’s not innovating on its own, it’s responding to prompts – those prompts come from humans.
Imagination is one of our most powerful tools, and it’s how brands can continue to stand out when the quantity of content out there grows and grows. So, while AI can’t come up with something truly innovative, when it’s recombining elements, it absolutely can spark some interesting ideas.
If we’re feeling like nothing is quite ‘wow’ enough and we want to try and kick start something completely different, playing with random ideas can really get the mind going.
As an example, I asked AI to come up with a new business idea that combines burgers and space travel. It gave me “Cosmic Burger Voyages,” with a list of features, including VR space tours in the queue and zero-G dining pods. That idea isn’t something I’m going to do anything with, but the act of combining two random elements to spit out something interesting definitely got me thinking.
Within any creative process, there’s a need to demonstrate your idea in a way that can easily be understood. Quick scamps and sketches will get us so far, but sometimes it’s more effective to show real visuals to get buy-in or sign off.
Using AI to mock up an idea saves a huge amount of time, especially when you’re presenting multiple options, ideas, or creative routes. Instead of a designer having to effectively do each full design, or spend hours making a composite of the elements we need to show, AI can spit out something pretty impressive.
You can get feedback on something more akin to what will be the final product, without the time investment.
Then, and this part is crucial, when you have an approved direction, you create the final design completely yourself. AI-generated visuals shouldn’t end up in final creative – using them opens up a lot of issues, including the fact that your design is not copyrightable and could be reused by literally anyone.
Then there’s the issue that a lot of generative AI has been trained on material that is already under someone else’s copyright, which could leave you exposed to legal challenge as well as reputational damage.
Using AI for creative endeavours naturally makes plenty of people nervous. The key to using it effectively and appropriately is finding the right places to activate AI and keeping it away where it’s not appropriate.
For example, within CRM, let’s think about where we would and wouldn’t want to use AI-generated content…
As the purpose of the subject line is to get someone to open your email, then the rapid optimisation enabled by AI is helpful. It can spit out a dozen versions and start whittling them down to the most effective as the emails are going out.
But the content of the email itself? That’s where you want the human touch – building relationships and prompting action is the purpose of CRM. Crafting copy and design that feel authentic, developing something new and engaging, and representing the brand flawlessly all take at least some level of human creativity and experience.
Effectively, we can use AI to augment creativity, taking on the jobs that are more time consuming or routine, leaving the humans more time to focus on ideation, innovation, and novel ideas.
Right now, AI isn’t coming up with the ideas that humans do, and it doesn’t know how those ideas can be developed to work in CRM. Giving our developers and creatives time to work together and brainstorm new innovations in email interactivity is the way we’re able to create emails that go beyond the expected – just check out some of our case studies for what we’re talking about here.
In 2016, 20th Century Fox used AI to create a film trailer for the horror film ‘Morgan’. The IBM Watson platform analysed hundreds of existing horror film trailers and then pieced together a trailer from the completed film.
It saved a lot of time – trailers usually take weeks. As a gimmick, it was great. It was a talking point, it got people to watch the film, it got press. In short, it did exactly what a trailer is supposed to do.
But that was the first of its kind. What happens if all trailers start being created this way? They’d all start to look the same and each genre would end up with a formulaic structure. Films would struggle to differentiate themselves – defeating the entire point of having a trailer. Eventually, the non-AI trailers start to become the ones worth talking about, the ones that surprise and entice.
The same issue would be true in CRM. If a particular email template succeeds, how soon will it be before all emails with the same objective start to align to that template?
How boring our inboxes will become – don’t forget, AI isn’t going to come up with a completely new, groundbreaking concept for an email. It’s going to amalgamate a bunch of existing emails. It’s not going to ask, “hey, can we have a fully interactive word search in an email?” and then spend the time to figure out how to make that happen.
Brands that want to stand out still need that human innovation, human curiosity, and human creativity.
Handing over creativity to an AI will always include compromise. The job of creatives is to figure out where that compromise is acceptable and where it isn’t.
The answer to that will differ for different brands, different businesses, and different scenarios. If your priority is having a large volume of content very quickly, then AI is going to exponentially increase your content production.
If your priority is having content that’s unique in its tone, style, and point of view, then you need human creativity and time.
The point is that AI creativity isn’t an all or nothing proposition. For any agency, client, brand, or individual creative, there’s a question to ask and keep asking as the technology continues to evolve, “where can AI help me?”
Creatives need to be on board with AI. There’s a phrase that keeps popping up when this topic is discussed – “AI won’t steal creative jobs, but creatives who know how to use it might.”
Businesses will be looking at the commercial opportunities from AI, and that will inevitably affect creative teams. If we’re the ones who know what the opportunities are, know what we need to do to get good results, and know where to avoid AI and the issues it can introduce, we’re the ones who will be leading the change.
Without knowledge, we can’t influence. Without curiosity, we can’t evolve. And without embracing and harnessing new developments, we’ll be the ones at their mercy.