By Jon Crowley, FUSE Create, Partner-Head of Strategy
The practical case for AI is that it’s an incredible tool for taking care of “grunt work.” Grunt work defined as things that are hard, unglamourous, but important. The entire category of grunt work is being redefined in our industry and beyond, and what we put in that bucket is going to reveal a lot about what we believe, at an individual level, a business level, and a cultural level.
What’s interesting is how many people who until recently would not have called core tasks of their job grunt work, are now treating it that way. In advertising, I’ve seen a lot of different groups quickly identify their own low hanging fruit. The grunt work of taking meeting notes and drafting a contact report (account service) or conducting a high level research audit on a new industry (strategy) or seeing if someone else had used this exact language / concept before (creative) or filling out timesheets (everyone) seem like perfectly reasonable roles for AI.
This gets more interesting as you explore more capable agentic solutions. Is all research grunt work now? Is all process work? Ideation? Synthesis? Competitive analysis? At what part of the process does expert direction at a step by step level become essential? Does it ever? Every person or company in every use case needs to figure out where the line is for themselves, which can be blurry when facing first hand how easy it is to make stuff with AI .
When you outsource things like writing, communication, relationship building, advice, editing, or design to AI, you are 100% gaining an efficiency advantage. You are also declaring those things to be grunt work. And if you are declaring the specific things that differentiate you, your business, or your product, to be grunt work, you are making the case that you, your business, or your product, are a commodity. Because fundamentally that’s what AI generated work is; unless you’re running a custom foundational model.
If you do not have a compelling argument for why people should pay you to pay for tokens to pay OpenAI, Anthropic, Google or X (or Mistral or Cohere or…) to generate the output, you are basically making a compelling argument for your role, business or product to not exist except as a thing someone asks AI to do directly, cutting out the middle man.
Grunt work is a commodity. Meaning unless you add something of value to it, it is ruthlessly competitive and easily replaced. Paying anything above the bare minimum for it is effectively a hedge against needing to train someone to do it again, and the exact same framing applies to AI. Which is why, in most cases, people and businesses have well thought out, clear arguments for why the things they do, are not just grunt work.
This is not an argument against using AI. It’s here, it’s pretty incredible in terms of what it can do, and you can use it to do things you wouldn’t have considered plausible due to affordability or timing concerns even a year ago. But it’s definitely an argument to consider what is actually grunt work and what isn’t, even if you can generate a passable facsimile of what a great and considered version would look like.
I don’t want the grunt work version of medical advice, or the grunt work version of a professional relationship, or the grunt work version of creativity. I’m already seeing some of the least engaging advertising of my life out in the world because someone who doesn’t know better—or in some cases, should know better– thinks it looks “good enough”. Good enough is fine for some things, the same way it was when I was stocking and facing shelves in a grocery store at 15. But the only thing good enough has ever made someone feel is sated, or indifferent.
That’s probably fine for grunt work. It’s a little unsatisfying for everything else.
