Subscribe
Share
Share
Embed
What are the best brands doing to stay relevant, build trust, and create content smarter?
At Share Your Genius, we have the same questions, so we're tapping the best in the space for their answers—one voicemail at a time.
Join us each week for quick hits of insights from b2b marketers and leaders.
[00:00:10] Amanda Smith: Every marketer wants content that feels real, right? Something that connects, converts and actually sounds human.And listen. I, just like everybody else, am using AI but when AI is a part of your process, that line between efficiency and authenticity can really blur fast. But I truly, truly believe that we can use AI and still sound human and authentic at the same time. So I decided to ask Nathan Thompson about it. He's the senior director of product marketing at copy.ai with a long background in content marketing and strategy. In his career, he's helped build a system that treats AI less like a magic box and more like a very smart intern. And he says, the secret isn't the prompt. Contrary to popular belief, it's the context. So when you train AI with your brand's story, audience and values, it doesn't replace that human touch. It just makes it stronger. So I asked Nathan, how do you make AI generated content still feel authentic? Here's what he had to say.
[00:01:36] Nathan Thompson: That is a wonderful question. I was the head of content strategy at Copy AI and recently switched into senior Director of product marketing,
[00:01:46] Our goal was to solve, go to market at a bigger level than a department. So if you had essentially like a solution for marketing with AI we think that that's great. It can also scale the silos within marketing that don't speak to sales and customer and product. We just wanted something that holds the data so you could get the right message to the right people across the whole function. So basically trying to unify the brand voice across the entire go to market motion.
[00:02:08] I think that every piece of content, and not just marketing, even that sales email that somebody's gonna get, you could call a piece of content. I think it always has four steps. You have strategy, you have first draft, you have polish, and you have publish. I only use AI for this draft piece. I don't use it for the strategy except for chat, maybe brainstorm, maybe find some ideas, but really like impactful AI is at this step. Then a human comes into polish, and then a human publishes. And I like to tell people, if you've seen really bad AI content on a website or on LinkedIn, that's still a human problem. It did not publish itself. Some editors said good enough. Clicked it.
[00:02:45] I think that a lot of people assume their issue is the prompt and they think I'm just not a prompt engineer. I'm not good at prompting. I think that's the silliest thing in the world.
[00:02:54] I would argue, Amanda, that you are probably selling yourself short in terms of your skills with AI and you are overcompensating what AI can do in terms of thinking that the issue's a prompt issue when really it's probably a context issue. And so you're probably prompting it really well, but your AI system might not know enough about your product's features, the benefits of those features, or the audience that those benefits would serve most.
[00:03:21] And so it's struggling to personalize content in the way that you want, and it's starting to sound like everybody else because everybody else is having the same context problem too. If you want really high quality and authentic content from AI my suggestion would be before looking at all of your prompts, to dive into your context and make sure that you have given it more than a day one intern, and you are treating it like it is something with a photographic memory, but really needs to be handheld through that onboarding process to understand your brand.
Alright, Amanda, thanks so much. Hope that helps. You have my number. Don't call me. I hate phone calls. You can text me anytime.