Leadership Lens: Artificial Intelligence

Q: We’re seeing more healthcare marketers turn to artificial intelligence to create things like page content and metadata.

Should organizations be jumping on the AI bandwagon or proceeding with caution while we’re still figuring this technology out? 

A: I’ve seen some of those “AI doom scenario” things where the AIs are generating so much content now that other AIs are taking it in as training data. Do we end up with this weird cycle where it actually gets worse and worse over time because there aren’t new voices coming into this? 

I think when we look at the hype cycle, we’re still very much ramping up that first hill. It’s clear there’s going to be some real value in AI. There’s a ton of experimentation going on out there, but the number of organizations where it’s been transformative at this stage is still relatively small. 

And I think we probably are going to go through a place where a lot of organizations make big investments, the majority of which kind of fail. Before we get to a stable place in the digital world, not just within healthcare, we really start to identify what the key use cases are that are going to be the drivers of meaningful value creation over the long term.  

I think people who, out of the gate, looked at these tools and said, “This is immediately going to replace a bunch of people,” I think are people who are most quickly climbing up that hill of unrealistic expectations. 

The tools are interesting; they can be really valuable, but I do think that everyone needs to figure out how they fit into their processes.  

Q: Where do you see AI currently making the most impact for healthcare marketing? 

A: We can talk about [AI] within the search context, I think that’s one of the places where it really is moving the needle in the most significant ways. The move from a traditional list of websites to go to to “Hey, here’s the answer you’re looking for,” is happening very quickly. Google is now rolling this stuff out; the move from Google being a search engine to being an answer engine has been going on for a decade, but [AI] has really moved it forward. 

It is starting to meaningfully reduce the amount of traffic that, particularly certain categories of websites, are seeing coming from search engines. I do think that’s going to change strategy around. We used to have SEO, now there’s kind of LLMO, large language model optimization for these AI models. 

It may change the way we think about content strategy. There are a lot of implications on that side, and that’s the place where we’re seeing the fastest movement. The number of zero clicks coming out of Google right now, where someone does a search and they never click outside of Google or they maybe don’t click at all, they’re getting the answer right there, those things have been growing and they’re skyrocketing now. 

We are seeing Google as a channel to feed traffic to our sites being in a particularly shaky position out there. I think [marketers] need to think about how they’re capturing traffic over the long-term, that’ll be a thing we’re going to be talking about a lot here over the next couple of years. 

Q: Should healthcare marketers be incorporating AI tools into their workflows today? 

A: Figuring out how to tailor these tools to deliver value at the right points within the lifecycle, if you’re just playing around with something, that’s fine.  

If you’re really going to look into doing that and producing commercial-grade, production-level code, it’s actually kind of hard to get that piece to do all those things for you. But within the life cycle of code, there’s a lot of places where AI can help support and handle different functions, and allow them to do a lot more work. 

But if that’s the only place where you’re looking to extract value, you may be missing out on where the biggest bang for the buck is. With writing content, I think people think about it like “I’m going to give [AI] a task and have it give me fully written pages of content.” It kind of can, but not very well, at least not today.  

Using it within the research part of the process, using it to give an editorial pass and give some feedback on things or using it to provide recommendations around, you know, SEO optimization, there are a number of steps that happen in those processes where you can train [AI] to be better than the actual writing of a page. 

I do think it can have a role and make us more efficient and improve the quality of output, but only in really careful ways. 

Q: Where is Geonetric headed in terms of AI use? 

A: We’re at this stage as an organization where we’re doing experimentation. We really haven’t built it into a lot of our processes at this point. I think we’ve been a little bit conservative about how fast we want to actually move this in anything that’s impacting real client work.  

The general generative AI world at this point, we are looking at it, we’re experimenting with it, but we’re also trying to be cautious and clear about not trying to get ahead of where the tools really are or our ability to use them consistently, reliably, and safely. We’re getting there, but we’re trying to move along in a way that doesn’t have the risk of compromising any of our clients’ work in the process of learning or making those things happen today. 

There’s a long way to go still on some of these tools and the ways that I think out of the gate, people thought we would be using them almost immediately. But I do think that between now and the end of the year, we’ll have a lot more to say about AI.