Tom says: “My additional insight for 2024 is actually to change your mindset in how you use AI to go with something that I call the ‘cyborg method’. Team up with AI but in a slightly different way than in 2023.

When AI was introduced, we had the ‘chat’ way of talking to AI. To be honest, we used AI a lot in the way that we thought we could use it – kind of like a search engine. We asked questions to retrieve data from AI, leaving a lot of space for AI interpretation.

The problem is that we got regurgitated content, and we could only produce some fluff and empty content. We hear about it a lot on LinkedIn, how AI is basically empty and shallow content. The problem with that, and how to fix it, is actually to feed AI with data because, instead of using AI as a search engine, I treat it more as a logical engine.

It's really great at analysing stuff. It knows a lot. It's really intelligent (I'm not saying it's not), but at the same time, if you feed it with the data and information and context that you want it to go after, it produces a lot better content.”

Understood. This is with an end goal. This is to feed AI with the highest possible quality data, the highest possible quality prompts, with the view for it to produce the highest quality possible content for you?

“Yes, but also do it in a structured way. For example, if we are a human and we're producing content, we're dividing it into specific structures. We are dividing it into steps. First, we're doing research, then we're producing a brief, we're doing an outline, then we're looking for keywords, then we're writing the article, then we're optimising the article for SEO, and so on and so forth.

There are multiple steps but, at the moment, I see that people expect AI to do it all in one step. It makes no sense. If we, as humans, can't do it in one step, why do we expect AI to do it in one step? This is the wrong way of thinking.”

Any particular AI we're talking about here?

“It doesn't matter. Whatever you use – if it's GPT-4, or Gemini 1.5, or Claude Sonnet, or Lama, or whatever you're using – treat it as a human. What would you expect it to deliver and what kind of things do you need to deliver to this human, so it knows exactly what you're after? How would you divide the steps for the other human to provide you with information?

If you think about AI in this way, you will produce much better things – and you will be able to get much more input from the AI on the things that you're working on.”

Do we have to prompt it the same each time? Because you talk about treating it as if it were human. Humans remember things, so you would like to be able to say, ‘Just remember what you did last time and do it like last time.’ Is AI at the stage where you can do that?

“No, you can't, unfortunately, because it's all based on how AI works. You have to understand (and this is not obvious to most people), when you chat with AI, every single time you answer, AI gets the whole discussion. It's not that it gets only the last bit that you wrote, and it remembers everything that is above. You're pushing the full discussion to it.

If you have a really long discussion, you will feel like it's not remembering what you talked about higher up. Yes, it doesn't, because the context window is limited. Every single time that you answer, you are pushing in all the information above so it can’t actually analyse that. It doesn't have this kind of memory in the way we think it does.

You have to build a profile of things that you want it to understand and remember, but you have to feed it with that every single time that you start a new discussion.”

That's a lovely way to put it, to build a profile of things that you want it to understand. This profile, can it live off the AI? Can it be some kind of online document that it can reference as part of the ideation process it uses to produce content?

“It can, but the problem is the technology that you're using. For example, if you were using Open AI technology, that would be the GPTs that you're building, right? You're feeding GPTs with specific knowledge that you want, but you don't have to use a GPT environment.

I'm a half-programmer, half-marketer. What I try to do, I use it from an API's perspective. I'm building my profiles of things on my side in my database. I just have to use it in a proper way.”

What would be an example of some elements that you would include in a profile?

“Brand voice. Simple as that. If you want AI to speak in a specific manner, you need to define your brand voice and you need to feed it to AI. With GPT, again, you have custom instructions that you can actually use for this specific reason, but in different AI models, there are probably different ways of doing that.”

What content is AI great at producing nowadays and what content does it need a little bit of a helping hand with?

“In my opinion, in all the content, AI needs a little bit of help. When you're producing content, the main reason why you're doing it is probably to either gain traffic or help your customers to better understand a specific topic. What AI is really good at is going after search intent, if you feed it with proper data. What AI is not good at is experience.

You, as a subject matter expert, talk to your customers and your product team, and you see how customers are interacting with your tools, services, or whatever you do. You can help out AI to understand how you feel. What's your subject matter expertise? This is exactly why Reddit blew up and probably quadrupled in traffic in the last couple of months – because there's a lot of expertise there.

In my opinion, you have to combine in SEO, and this is what Google says in their policy about what's spam and what's not. You have to add value, and adding value means you have to have the expertise. We're good at expertise as humans. We, as I said, talk to customers. We have a lot of experience that AI doesn't have.”

You talk about the fact that this cyborg method, as you call it, adds more value to content than you could add before. Is this because you're combining the power of humans and AI, so you're not just relying on one?

“Exactly, because we have to feel that we have superpowers and AI has superpowers. When we combine it, we can actually create something much better, much faster – and get better results.”

Will we ever be at a stage in 2 or 3 years’ time where we can just leave content to AI and focus on something else?

“Hard to say but, right now, a lot of people are trying to say that AI is not there, and they can create better content. If we still focus on thinking like that, I think we will lose our jobs. That's my opinion.

But, if we change our mindset into a way of thinking about how we can improve what we're doing by combining our powers together with AI, then we're safe – and I think we're important in the process.”

You talk about doing this at scale. What kind of scale should we be targeting?

“The question is, what do you want to aim for? The most important part for me is to actually have KPIs that I want to go after. Creating content is just a tool to get there. Depending on what you want to achieve, you will talk about the scale.

If you feel that you want to quadruple your organic traffic and improve the number of leads, let's think about exactly what kind of steps you need to take to achieve that. When it comes to AI, nothing changes from the strategy perspective. It's just a tool. Use it the best way you can.”

Is it possible for brands to compete with this cyborg method or producing content using AI and just produce content using humans?

“It's really difficult because the ROI doesn’t add up. You can significantly improve your ability to create quality content using AI. If you think that you can compete directly using pure humans, I don't believe so. Maybe you have different KPIs that you're after – again, choose the best method that leads you to the KPIs that you have.

If you use AI, that's fine. If you don't, and you feel that humans can bring you better results, do it. But for organic traffic, honestly, I don't believe that you are able to compete with tools that use AI. Because it's just much faster and much better. AI can go through much, much more data than you, as a human. I can't compete with AI in reading 300 pages of text.”

Do search engines not detect that the content has been produced by AI and are less likely to want to rank those articles highly and give them traffic?

“This is a question that many people ask themselves but, in my opinion, this question is wrong. In my opinion, people should ask themselves, is the content that I'm producing more valuable for the end user? How you do it doesn't matter. What Google wants, and says in their documentation, is the highest value to the end user.

If you don't deliver value, you're creating spam – and it doesn't matter if you're creating spam using humans or AI. They will deindex spam from the internet and, to be honest, I really love it because, in the last few months, Google has actually been doing a great job with removing a lot of spam from the internet, and it's making a lot of space for valuable content.

I haven't seen such growth in the last decade. I think, in around 2012, I was seeing such growth – when I was spamming with backlinks that were basically black hat SEO. Google couldn't catch it at that point. They came to a point where they were able to catch it. But with AI, at the moment, it's according to Google’s line.”

You talk about creating valuable content. One way to measurably create valuable content is to have some authoritative author associated with that piece of content. Is it reasonable to produce a piece of content in combination with AI, in conjunction with AI, and then to say that that piece of content has been written by a famous author in that particular niche?

“I believe that, if you're having valuable input into the article, you don't have to explicitly say that you were using AI to edit it or add additional things because it's like saying, ‘I used a pen for writing this article’, right?

At the moment, for me, it's an engine that helps me to use the input that I'm giving it to structure it in a way that is digestible for the user. Of course, if AI is purely writing the whole thing, then you should state that it’s written by AI. This is exactly what Google says. But I think everybody right now is using AI, in one way or the other, to edit their content. If you're using Grammarly, you're using AI.”

Good point. What's your strategy, then, for deciding what content to focus on? Because to a certain degree, AI opens up the opportunity to write almost limitless articles within our chosen industry sector. What do we focus on?

“It's all about finding the way to structure your content calendar. It doesn't have to be connected with AI because what I like to do (I'm a data-driven person) is I like to find keywords that reflect what I'm going after.

When I have my funnel (TOFL, MOFL, BOFL), I like to have a mix of all the things together because, if I only create bottom-of-the-funnel content, I don’t feel like I'm being treated as an author who talks about the specific topic and has value when talking about the specific topic.

You have to go after bottom-of-the-funnel, middle-of-the-funnel, and top-of-the-funnel – even if top-of-the-funnel is not converting at all. This is how you create your authority profile, but it's all about data. Use keywords that your competition is after. Add additional trends to the content that you want to create. Use AI to find gaps that you're missing in the content that you already have. The options there are limitless.”

You've shared what SEO should be doing in 2024. Now let's talk about what SEO shouldn't be doing. So, what's something that's seductive in terms of time, but ultimately counterproductive? What's something that SEO shouldn't be doing in 2024?

“Basically, in my idea about how SEO is working and SEO experts are working, they should focus less on keyword stuffing and focus more on search intent. At the moment, when we see all the ranking factors and all the ranking tools showing you how well you did, and there is a graph showing that you're at 70/80%, they're mostly focused on keyword stuffing.

They just say, ‘Add these keywords 2 times, add this keyword 3 times, and you're good.’ From my perspective, the question is, did I nail the search intent? If I add these keywords, will it make this content more useful and more valuable? It's as easy as that.

I'm not saying it's not working at all. I'm saying that, according to Google's policy, keyword stuffing is against their goals. It's written right there. Just Google their policy when it comes to content and read it through. What I'm saying is, focus on the search intent. Focus on the things that people care about and would like to see in your content. When you do it, it will work wonders.”

Tom Winter is Chief Growth Officer and Co-Founder at SEOwind, and you can find him over at SEOwind.io.