Stay ahead of the game by embracing AI

Nitin says: “Embrace AI and stay ahead of the game.

AI is becoming more and more important for SEO, and any other marketing channels for that matter. It gives businesses the tools and insights they need to stay competitive in the digital landscape. At Botpresso, we’re using AI for data analysis, content production, optimization, translation – you name it. We do everything with the help of AI, and it gives us 10 times more efficiency. We’re also writing a lot of Python scripts to solve complex SEO problems.

Businesses can leverage all of these opportunities and much more. AI is evolving, and you should embrace that.”

Do SEOs need to use AI to make themselves more efficient?

“Absolutely. In whatever you’re doing right now, you will see patterns. Wherever you see patterns – in data analysis, producing content, translating content – you can think about using technology. It brings a lot of efficiencies – and it’s cost-effective as well.

If you’re paying a content writer 100 euros to write a content piece for you, you can probably do that at half the cost or less with the help of machines like AI. It’s also efficient. You can produce a lot of content in a single click, which could take months or even years to create manually.”

What is the best data to analyse with AI, and what software do you use?

“If you want to compare your crawl data to that of your competitors, you can feed it into an AI tool. You can train the AI to understand that data and give you insights. This is the main thing we are using it for. You can give the AI that crawl data and ask it to identify the differences between them and the strengths of one brand versus the other. That’s the data analysis I’m talking about.

However, you feed in any kind of data. You can feed in data about the events that might be affecting your seasonality and ask AI to give you a forecast, for example. For everything that used to take a lot of time, where you were looking at hundreds of parameters, you can now just feed that data into an AI tool to help you understand what is happening and how it could affect things. It does the magic for you.

Currently, we use a lot of software for different applications. For data analysis, we prefer to write our own Python scripts because that gives us flexibility. Right now, there are people in my team with zero engineering experience who are all engineers with the help of ChatGPT. They’re all writing Python scripts to solve complex SEO and data analysis problems.”

How much involvement should AI have in the creation of content?

“I have been doing this for many years now. AI is still new, but I started content production at scale 6/7 years ago. I started with normal templates, then conditional templates, then we used a machine called RosaeNLG (a natural language generation machine), and then we started using more advanced technologies.

Now, ChatGPT is definitely my go-to tool when it comes to content production, but I’m also using some more sophisticated and advanced tools like Jasper and Writesonic. These tools are great if you prefer a nice-looking UI and a more structured way of producing your content.

I’ve been using AI for content translation as well. I have been using DeepL, which is a brilliant translator that’s 100 times better than Google Translate. Now, we are also using ChatGPT because it gives the flexibility to train the machine. I can give it the content and tell it that I need a more expanded version, along with the translation. I have everything in one place. However, it does need some manual intervention to train the machine and control the outcome.”

Do you still need to have human involvement in that content creation process?

“You definitely do because, at the end of the day, it’s a machine. The machine doesn’t understand a lot of things, like your brand guidelines. It is just generating content based on whatever data it has processed and spinning another version of that.

Ethical considerations are also very important. For example, I’m living in Germany and data privacy is a big topic here. AI can give you a lot of sensitive information, and how you use that can create problems.

On top of that, when you are writing this content, you are not only writing for machines. Gone are the days when you were writing content for the machines and that would make Googlebot happy. Now, the content should be more user-centric. If a user is happy, you don’t have to worry about making the search engine bot happy, because they will be happy already.

You definitely need to include personalisation and a human touch – and ethical considerations and data privacy are also part of that. To do this, we have someone who understands the product we are writing content about sitting on top of the process and controlling everything that is generated with the help of AI.

It’s all about the initial quality of your prompts, ensuring that the content is sticking to the brand guidelines, and making sure that the writing style is correct to begin with. Then, you should have a knowledgeable specialist to review that content and tweak it before it is published. We published an ebook on this recently (which is available at Botpresso.com/AI-prompts-for-SEO). It covers a lot of the great prompts that are working like magic for us, and much more besides.”

A direct translation often won’t read fluently to a local audience, and it can miss important context, so should a local specialist review that content as well?

“Again, you should follow that same process. There should be someone involved who is a domain expert – a native speaker, who understands how people in that location think and communicate about what you are offering. That person should be involved in the whole process, from the beginning. They can then tell you that what you would call buses in Germany are sometimes called coaches in the UK, and football in the UK is a different game altogether from football in the US.

With the help of a local expert, you can train the machines to understand these nuances. Then, your machines will produce or translate content that is of a much better quality than it would be without that training. Once the content is produced, the domain expert can take a look at it again and give their final approval.

When we were using DeepL, it was working pretty well for most European languages. Interestingly, though, there were languages that it was not as good at. English to French translation was not that great, and neither was English to Swedish. For English to German, English to Italian, and English to Spanish, it was close to perfect. The native speakers who were reviewing the content were happy and they were hardly making any changes.

However, our French native speaker was disappointed by the content that DeepL produced in the beginning. We trained it, and we got better at the process as well.”

How are search engine algorithms evolving with AI, and what do SEOs need to do to take advantage of this?

“The SERP has been changing forever. I can’t remember a year where we didn’t see significant changes in the SERP, and that will continue to happen. Now, with SGE, users can ask for something (particularly informational intent content) and get the answer in the SERP itself. They don’t have to click anywhere.

I think this is more of a problem for search engine bots, and how they show this content, than it is a problem for SEOs. They want to make money and, if they provide the information without the user having to click anywhere, how do they make money? They want to keep the links there – both the paid links and the organic links.

The businesses that are serving more informational intent, like blogs, will probably suffer because most of their content will be available without any research. I used to go and do detailed research on things like what the maternity policy for a company should look like, or how the RICE model of prioritisation works. Now I just go to ChatGPT, give a command or two, and get the answer I’m looking for. That makes life very easy. Before, I would probably have been given five links to different websites but that has now reduced to zero.

However, when it comes to transactional queries, I’m still going to Google and searching and transacting in the same way that I was before.”

Is it still important to publish blog posts and pages targeting long-tail queries to gain that initial traffic?

“I think you should definitely still be doing that. Look at how the SERP is changing. Along with the detailed answers that the search engines are giving, they’re also providing links to where that content was extracted from. In that case, you will gain brand visibility. If someone is searching for something and the answer appears, and it was covered in your blog, you can be mentioned there. People would know that they got that answer and it was covered by you. That is brand visibility.

Also, if you’re afraid of what is happening with SGE, you can block the AI bots from using your content. Right now, I wouldn’t advise doing that for transactional businesses, but more news-centric businesses might consider it.

Currently, I’m working with a business that talks about everything to do with startups (funding, news, etc.), and they have a lot of content that is behind a paywall. If all of that content is accessed by AI crawlers, and it’s being pulled and published for free, then their revenue model will be compromised.”

If an SEO is struggling for time, what should they stop doing right now so they can spend more time doing what you suggest in 2024?

“Start with the basics and try to understand how AI works. Try to understand different types of prompts (and you can find all the information you need in my book). Then, you also need to play with different machines that can help you do what you want to do.

ChatGPT is pretty awesome, and so is Perplexity AI. Then, if you want to go with more sophisticated solutions, you can try Jasper or Writesonic for content production.

It’s important to understand the SEO use cases that you want to try. You can’t just get comfortable with ChatGPT; you need to understand what you want to solve with it. Once you have those solutions, then select the AI tools or platforms from those that are out there, and the hundreds of new platforms that are coming. Try those tools, train the machines, tune them, and test them out. You won’t get the ideal outcome on the first try, so iterate and refine.

Then, it’s very important to keep ethical considerations and data privacy in your process, especially when you’re producing content. A machine is producing content for you, so it will not think about your user. They do not know what users are looking for. Also, those machines might produce some content that you might not want to show on your website for data privacy reasons. Make sure that is also covered.”

Nitin Manchanda is Founder and Chief SEO Consultant at Botpresso, and you can find him over at Botpresso.com.