With AI Agencies and their clients should use the early iterations of AI to experiment, fail and experiment. Most of all, however, they should use it to satisfy individual and collective curiosity and drive positive change, writes Louis Keegan.
When broaching the subject of generative AI and its use within agencies, opinions differ wildly. Is it the greatest creative opportunity of our time or will it have a much lesser impact? It’s been said that we are underestimating the long-term role and impact AI will have in agencies, and I’m inclined to agree.
In the same breath, there is no ignoring the vast amount of headline grabbing “investments” we are seeing. Supposedly, vast amounts of resources are being poured into this technology by agencies across the globe. When any new technology or channel presents itself, agencies race to position themselves as being thought leaders at the forefront of understanding and AI is no different.
I don’t doubt the requirement for this investment as agencies and clients seek the best way to adopt this technology. I do however approach the language used around these claims from a place of inquisitive scepticism. To quote the American AI researcher Eliezer Yudkowsky “by far, the greatest danger of Artificial Intelligence is that people conclude too early that they understand it.”
AI is being used in agencies across the globe at varying degrees of implementation and experimentation as organisations explore the capabilities as well as the limitations. This raises the very real issue of expectations vs results. Anyone who has experimented with this technology will have noticed outputs are often far removed from the glossy demo videos you see across social media. That isn’t to say you can get great results, it just takes a lot of work behind the scenes to get there.
Pilot & Co-Pilot
This is where I feel the pilot/co-pilot relationship model (If you can call it that) really comes into play. AI and its use within agencies has to be done with human first approach and when experimenting or using this technology individual oversight and responsibility are key in driving results that meet the standard. This is most certainly easier said than done. AI is being touted as this time saving device that will free us from the burden of office work and allow more time for creative ideation. So, when AI isn’t delivering on this promise, frustration rears its ugly head and this is what I think is AI’s biggest hurdle currently.
If agencies really want to adopt AI in any meaningful way they have to account for this. There needs to be an understanding that this is a journey, and we are taking our first tentative steps along that road. Agencies must give their people the space to fail as one can be certain that mistakes will be made along the way. However mistakes lead to understanding and understanding leads to change.
In-house Strategy
Experimentation and play are consistent themes that I’ve encountered when chatting with industry colleagues around the most successful ways of building an inhouse AI strategy. Allowing that space and fostering a sense of individual ownership and responsibility are key in driving meaningful and lasting adoption of this technology. At its core this is what AI adoption should look like, when we look past the headlines of huge monetary investments. It’s about creative exploration and the people that drive this change.
Agencies that go into this with the attitude of expecting instant results are doomed to be frustrated and dissatisfied and the media has perpetuated this notion.
When looking to build an in-house strategy, look to your teams, review what’s already being done, many are already using AI tools privately and you could be surprised how! Discuss how it’s being used and how you can scale it, find areas of the business that can be optimised.
For the strategists it could be research, semantic based LLM search tools are incredibly useful research assistants. For the designers maybe its visual mood boards where generative visual tools are brilliant at capturing tone or feeling.
The idea is that you and your teams need to start playing with these tools, build an in-house use case, maybe it’s a spec project or just some examples that demonstrate an understanding of the technology and its limitations.
Then I would suggest starting that conversation with your clients, show them what you have done and what you could do for them, go on that journey together and work collaboratively. This is the most effective way I’ve seen AI onboarded within organisations. It’s not about the massive organisations pumping vast amounts of resources into a tool, it’s about individual and collective curiosity and the passion to drive positive change.
Louis Keegan is a Strategic Account Manager with Pluto.
