Incorporate context to improve your AI-powered content

Erik says: “Use AI across the entire SEO process. In particular, we’ve only seen the beginning of what AI content could be.

A lot of SEOs and content marketers have a preconception of what AI content is. There’s a clear divide between the first phase of AI content that we’ve seen and what you could be doing. The first phase is ‘recycled’ content because the large language models are using the knowledge they already have to create a piece of content that isn’t actually new.

If you think about what Google and the other search engines want, they want unique content. You should use context in your AI content creation because context is what will separate recycled content from ‘amplified’ content, which brings context into AI content creation to make it unique.

The context could be a human layer, which is where a lot of SEOs are seeing good results. You use AI tools to speed up the process, but then you bring in that human touch to give it a unique aspect. You can also bring in context by including a PDF or other file in the content creation process, which significantly improves the capability of an AI tool and allows it to interpret information in a whole new dimension.

We have a customer in the travel space who had a deck plan for a cruise ship. If you want to go on a cruise, you use the deck plan to figure out where you want to stay. However, as an image, it wasn’t very accessible. As SEOs, we like things to be indexable.

Using Wilgot, our AI-powered writing tool, they took that image and turned it into a series of perfect content marketing articles talking about where you want to stay depending on the location of the swimming pool, restaurants, different decks, etc. The AI could understand the context of the ship and how each cabin relates to the other. It would have taken weeks for humans to write those articles. That’s a good example of using context.”

Should you use AI in technical SEO, for researching and locating issues on websites and for repetitive tasks like creating titles and meta descriptions?

“Yes. I think we’ve only seen the beginning of the beginning with what AI will do to enhance all of our processes.

If you think about it, technical SEO is especially prone to automation and amplifying the work you’re trying to achieve. We haven’t seen much of it yet but, in 2025, we’re 100% going to see more AI-assisted tools for technical tasks.

Coding is one of the use cases where AI is really taking off, with GitHub, Copilot, and so on. Marrying coding with data analysis is a sweet spot for large language models. There will most likely be a bunch of tools released in 2025 that can help you.

There might already be tools out there that I don’t know of, but things are going to get far easier when trying to fix basic issues like meta titles, and 404s. All the grunt work that was previously very manual can now be shifted over to AI for the heavy lifting.”

If an SEO doesn’t have much collateral available to them, how do they establish the context within the content that is created by AI?

“There are different angles to it. For one, you can apply AI to give you more capacity around keyword research. An AI tool can understand the context of your website and think as if it were a customer trying to buy your product. It can then think outside the box and find ‘blue ocean’ keywords.

The most common way of doing a keyword plan today is looking at 5 competitors and then copying whatever they’re doing to try and rank for those keywords. However, there is an entire ‘blue ocean’ of keywords that no one’s looking at, because search behaviour is changing every day, month, and year. AI has the capability to look beyond that.

A more short-term strategy would be, if you have an existing article, AI can create 10 versions of that article by using the same information as a base and then expanding on it.

For another example, I held a presentation at brightonSEO, and I took that presentation and instantly turned it into 10 perfect content marketing articles, aimed at different searches, with different themes. You can start with just one asset; it doesn’t have to be much. You can just summarise your unique way of thinking about a subject.

The key is that you have a unique point of view that no one else has today. You can use that as a base and then expand it to get different angles on this piece of content. It doesn’t have to be a 100-page PDF. It can be 3 pages of your thinking that the AI can then tap into and expand on.

Video and audio transcripts, like podcasts, are perfect because they are locked into one format. Most people don’t have the time to listen to an hour-long podcast, but there’s a lot of good thinking there, so it’s a perfect use case to turn that into a series of articles around the topic.”

Is there a step between asking AI to think like a customer and having it create content for them? Do you ask AI to come up with keyword phrases that are relevant to that particular customer?

“Yes, you can – and there are different ways of doing it. You can use ChatGPT and then combine it with other existing keyword tools. Our tool does it out of the box, and that’s one of the big potentials with AI across the whole SEO process.

In keyword research, there’s so much undiscovered search behaviour that no one is addressing because of the way that the system works. People do keyword research, and they only look at websites that are competitors, and then see what they’re ranking for. It’s been difficult to spot new, emerging search behaviour, but the potential is there.

I’m looking forward to when SEO can become more real-time, and focused on the trends that are emerging in how people are searching – similar to how social media works, and how a hashtag can go trending. That doesn’t happen in SEO because the system is so much slower. You have an agency doing keyword research on day 1, then they revisit that keyword plan a few months later. In the world of consumers, a lot happens over a few months.”

Where’s the sweet spot between relevance and volume when it comes to keyword phrases?

“We’re moving much further towards having a share of voice in terms of topics, especially with AI overviews. There is the question mark of relevance with the whole concept of keywords.

Keywords are what the user inputs, so they will always be there as a foundation. However, over the last couple of years, we’ve seen Google moving more and more towards topical relevance, where one keyword might generate X number of searches, but there are relationships that we can’t see that Google knows about.

They know that a topic has a number of related subtopics. If you’re talking about one of them, you could get impacts from another that you didn’t know about. That’s why keyword research is still relevant, but it’s much more about topical authority and topic research – and gaining that trust so that your brand is allowed to talk about a topic.”

Is there a point where someone could be using AI too much?

“The number one thing is whether or not the content is helpful, relevant, etc. The standard Google playbook still applies. With large-scale content creation, the key question is, does it add value to the SERP? Is it unique? Is it bringing something new?

Large-scale content creation can work because there’s so much knowledge and data locked up in different formats. There are so many companies that have billions of PDFs sitting on desktops and knowledge that needs to come out. The only way to tap into that would be with AI. If you apply it in that way, where the AI amplifies an existing dataset, that’s great.

If you’re generating hundreds of thousands of pages that are just reusing information, you’re not adding anything new into the mix. You’re not going to add value to the users. If you do that, you’re going to get some short-term impact but, compared to the impact of something unique that’s either created by a human or amplified by AI, it’s going to be magnitudes of difference.

We’ve only seen the beginning of this trend because the capabilities are moving at such a high speed. Some larger companies might be doing this internally, but it’s very far from reaching the true potential it has.

Right now, it’s amplifying your work; not doing all of the work. That’s where we are for the foreseeable future. Over time, the capabilities will increase, and more work can be handed over to the AI. For now, you need the human touch with AI content, to add relevancy and compete with others who are using similar tools.”

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 2025?

“Start applying AI to the tasks that you do manually but could be done by AI.

It’s very early days for AI keyword research, for example, but having it do the same sort of practice – looking at competitors, getting their keyword set, and matching it with yours – is something that AI can do a lot quicker. Also, for the basic idea-generation tasks in your workflow, you can use AI to help you expand on a topic and see what other angles you could add to a piece of content.

It can also help you get away from writer’s block, when you try to think of something to write and your mind is blank. AI is fantastic at that.

One of the challenges with using AI to help you come up with ideas is that those ideas are based upon content that already exists, so it’s less likely to be original blue-sky thinking. That’s why you need context.

If you have context of any kind, that’s how you can get the most leverage from AI. If you just ask it for ideas, it’s going to give you something that’s already out there. If you give it something unique, even just your personal thoughts in a PDF or a Google Doc, you let the AI tap into your brain.

That’s where you’re going to get most of the impact; when you bring in something unique that the internet has not seen before, then you build on it.”

Erik Wikander is Co-Founder and CEO of Wilgot, and you can find him over at Wilgot.ai.