Kathleen Booth builds on Shopify's "Can AI do this?" hiring framework by adding an ethical filter: "Should AI do that?" She walks through her product-marketing workflow at a startup launching 12 products in 12 days, using voice dumps and custom GPTs for first drafts, then human editors for accuracy and brand voice. AI will eliminate some roles, but she sees it as freeing people for strategic work.
Chapters
It was about a year ago that I think Shopify CEO Toby Lucke came out with a memo that was widely circulated where he was like. And it was specific to hiring, but he was saying, when we go to hire, the first question we're going to ask is, can AI do this? And if the answer is yes, we're not hiring a person. And it was very controversial at the time, but, you know, and I was probably one of the people who was like, ooh, that feels kind of icky. But the more I've kind of, like, traveled along this journey, the more I've come to believe that that is actually the right way to approach it. Which doesn't mean that I don't believe we should have people and that I'm, you know, about all about eliminating people from the equation. I think what it just means is that, you know, like, we've got to be smart about how we grow. And. And I would use this analogy of, like, over the years, we've had other technology that has introduced efficiency like this. Not at the pace that AI has, but, you know, we think back to computers and, like, things like Microsoft Office and word processing and spreadsheets and stuff like that. And, you know, there. That introduced a lot of efficiencies that probably eliminated roles, but it just happened over a much longer period of time, so it wasn't as obvious. And do we, you know, in the age of spreadsheets, do we still want people manually doing all of that work? No, we really don't. It's not smart to do that. But, you know, and so I think to me today, the question is really, like, when I think AI first, I think, can AI do that? But I would introduce one caveat to what Toby Luecke did, which is the second question is, should AI do that? And so AI can do a lot of things. You know, it can write, it can do data analytics, it can ideate and help you with strategy. There's just, you know, it can. It can create automations and workflows. It can do tasks. But you do have to ask yourself, should it do it? And, you know, so some examples are like, right now I'm leaning very heavily into AI for product marketing. And I'm probably gonna start to write a little bit more about this soon. But I kind of have to, because we're doing, As I said, 12 product launches in 12 days. And it's just an insane pace. And we're a startup, so I don't have a huge team, I don't have a product marketer. On my team, it's me and like one other marketer and a BDR and you know, the rest of the company, which is like the product team, the CS team, et cetera. So for us to do this, I have to use AI. That's the only way to make it happen. You know, the question is like, to what extent should AI do it? And I think you alluded to this where you were like, they, you know, there are some people who will just take whole cloth, like this is something AI put out. I'm just going to copy it, I'm going to paste it and I'm going to go. I think, you know that product marketing is a great example of where you like can't do that. Like AI has. I've trained it really well. I've gotten pretty good at like training my custom GPTs and things. And yet it still does hallucinate and it will spit out fully baked product blogs that say, like, this feature is available to all customers. When I'm like, it's actually only available to enterprise customers. So like you have to be really careful and not just about the factual date things, but also the that you're capturing an authentic human voice and tone and that your humanity shines through. And so I'm a big fan of like first drafts with AI and then have a human go in and like update it and make it reflect, you know, their opinion, their voice, their tone, their insights. So it's actually to back up my process with like generative AI is to bookend it. I start by brain dumping into a voice assistant and telling it what I think and giving it my point of view. I then use that to train the AI. The AI writes a draft and then I go back and kind of massage it so it done properly. The end result should be something that reflects my point of view because that was the starting point. So that's just an example of how I'm thinking about it. But I do, like, some people may disagree with this, but I do think that there are roles that will be eliminated because of AI. And I think in many cases that's okay. But on a personal level, like I didn't want to be involved in going to a company where I had to fire people because of it. Like, that doesn't feel right. And so that's part of why I went to a startup where I have this opportunity, opportunity to build from scratch in an AI first way. So it's not about job elimination, it's about like when we hire, we're going to hire people where we really need humans, and we're going to have all of our humans working on the things that AI can't do so that they feel like their brains are being used for, like, the really strategic things, which I think is exciting, and it's, like, makes your job more fun when you're not doing the menial stuff, so.
