Turn AI from a problem to a partner

Giulia says: “Turn AI into your partner, not your problem.”

How do you do that?

“The industry as a whole has been panicking a little, ever since the AI revolution started taking over. There's been quite a lot of conversation about AI taking over our jobs and replacing us.

I don't think we're there yet. I don't know if we'll ever get there either, because I think we have some aces up our sleeves. The way that you turn AI into your partner is by knowing exactly where to use it versus where not to use it.”

What aces do humans have up their sleeves?

“We know that AI is really great at picking up content, summarising it, and finding insights, but it still lacks context most of the time. That’s something that we need to provide with the data sets that we're analysing.

The most powerful ace up our sleeves, as humans, is the fact that we can actually tell when we're seeing AI versus when we're interacting with a human. There have been many studies that have shown that. When we know that we're interacting with AI, we're much more diffident. Whereas, when we don't know that we're interacting with AI, we tend to try and find those human sides of AI to be able to connect with it.

I've done a study to figure out whether humans have an edge over AI, because I was hearing so many conversations about AI taking our jobs and humans not being able to recognise it. I'm a woman of science. I want to figure out for myself if that's the case, because for me, I think the human edge will always prevail.”

Do you think there will be a time in the future when humans won't be able to tell that they are interacting with AI?

“I think there might be an age where we won't be able to tell, but at the moment, we're giving so much credit to AI, but not giving credit to ourselves.

Historically, we have a very good track record of adapting to new revolutions and using them to our advantage. I'm not saying that's always going to be the case, because we've also seen history repeating in ways that are not great for humanity. However, neuroplasticity is our human superpower: our brain adapts to what we expose it to. The more we're exposed to AI, the more discerning we'll be of it, especially if we're exposed to it every day.

This is something that has been noticed in other studies as well, and something that I've banged my head on when I did my own study. I thought I was going to see a human detector for AI, and we would have a different reaction for AI stimuli versus human stimuli, based on the fact that we detect AI stimuli as something strange that we're not used to.

However, what I noticed was something opposite. It was a heightened response for the human component. When I then went to try to explain it, I found literature discussing the fact that, currently, we're not able to replicate the studies on the uncanny valley that I based my entire experiment on.

The uncanny valley is the negative feeling that we experience when we see something that's almost human, but not quite. More recent studies have shown that we almost don't have that same response anymore, because we're so used to something imitating human life and human appearances. Now, it's the subtler differences that we use to distinguish something as being non-human.

Ultimately, I think we're always going to have an edge. Although the progress is very fast and we might be a step behind, we have the capacity to adapt to it, learn, find new patterns, and respond to these new outliers.”

What are the current limitations that AI has?

“We are still seeing all kinds of hallucinations. Where I've seen the most limitations, though, is where it doesn’t include context as part of the output that we get from it.

For example, I was recently talking to somebody who will be speaking at brightonSEO, and they were walking me through their decks. It was about using AI to create an intent detector, which essentially brings together all of this data from the SERP and from internal databases to find out what the intent of the user is.

What they explained to me was that the prompt needed to be tweaked several times to get something that was usable because they needed to put their own expertise into it, with the intent categories that they wanted to extract, etc. There were a number of human inputs that the AI could not achieve by itself.

Another limitation is the ethical implications of the misuse of AI. I don't think a lot of companies are thinking about this, because they're blinded by the scalability potential, but it has massive implications – not only because it's not right to be deceptive and create fake influencers to tell your story, but also because humans are naturally more diffident towards something that's AI-based.

It's a double-edged sword. Businesses that use AI to scale human-generated content by faking it are actually damaging their own brand and their own authenticity, which is what humans crave in order to connect with a brand and purchase from them.”

Do you think that AI will take some jobs in the future, unless SEOs change the way that they do things?

“I don't think we're going to be there in 2026. At the moment, we're almost at the peak of the adoption curve. A lot of businesses think that they can replace employees with AI. Obviously, some functions will be massively affected, because we've seen that already (for example, sadly, with copywriters and copywriting agencies).

When we talk about SEO, that's happening more with the procedural functions and the big data analysis that’s part of the day-to-day. Those functions might get affected, however, our role as SEOs has been changing for the past year and a half already. We should already be on a trajectory of change, where we evolve beyond what traditional SEO has always been about.

We've done it before. SEO 5 or 10 years ago was not what it is today. We’ve moved on from only caring about keywords and intent to now looking at the entire search journey, which is where SEO is going to go. It's no longer about Search Engine Optimization; it's about Search Experience Optimization, Search Everywhere Optimization, or whatever you want to call it.

There are so many channels that are part of the organic trajectory now that we need to take into account. While some functions might get replaced, there is still space for us to create our own path in the industry and tackle these new challenges around being present on other channels that are not necessarily Google, and the evaluation process and continuous exploration that we've seen search behaviour evolving into.

Maybe I am a bit of an optimist, but I really want to believe that, while some of our day-to-day might change as a function of having AI at our disposal, there is a new place for us in the industry.”

What areas of SEO should we definitely be using AI to replace at the moment?

“As I mentioned, AI is great for looking at data sets. What previously took a week of work now takes very little time because we can ingest great amounts of data, get insights out of that data, and with a few tweaks, get what we want out of it.

It’s useful for anything that’s more on the database side. For example, if you wanted to see the similarity between topics, the similarity of intent, and how this plays into creating new internal linking modules for an e-commerce site. It provides a lot of opportunities that were previously only available to a few people, with specialist tools that would analyse the entities of each page and identify the similarities. Now, you can literally do that with any machine learning model.

There’s so much more that we can take on these days. It might not be part of traditional SEO, but AI can help with anything that streamlines the journey, brings the user closer to us, and removes friction – like product descriptions that are based on specifications.

Obviously, you should never write all of your content with AI, but if you’re in e-commerce and you have a lot of product specifications that are just sitting in the catalogue, you can feed them into a machine learning model and create content that's actually helpful for the user.

In 2023, I did a study with a colleague of mine, Myriam Jessier, where we looked at which parts of a product page were most important when it came to people actually deciding to purchase something. The product description came up on top, particularly the specifications. Sometimes, they're a goldmine that doesn't really go anywhere, but I've seen them being featured as part of the AI overviews now.

Before, I didn't really pay much attention to the product descriptions themselves. If you have bullet points on the product page, you would assume that they were all good. However, when they're brought up as helpful content, then they might give you an advantage in being shown on AI overviews.

It’s also worth considering AI summaries of real user reviews. We’ve seen that on Amazon, and many other e-commerce sites are doing it as well. People obviously want to see reviews, and they want to see them from real people, but it can be overwhelming to go through each one to see all of the different pros and cons, and high cognitive load tends to produce decision fatigue. AI can bring in the best pros and the worst cons of a product, and people can make a decision much faster.

Our challenge, as SEOs and marketers, is the fact that attention spans are declining, and there are so many competitors that can take your user at any point of the journey. You want to reduce the friction to get what they want and what they need to convert and trust you.

Those are the areas where SEO can really embrace AI: anything that provides more information, facilitates navigation, and streamlines search journeys.”

Giulia, what's the key takeaway from the tip you shared today?

“Don’t panic. Get familiar with the technology available, but know where to draw the line.

There's so much more that we can do with AI, but bear in mind that there is also a fine line between where we should and shouldn't use it. Not only due to the ethical concerns, but also because it can negatively impact your customers' trust towards you.”

Giulia Panozzo is Founder at Neuroscientive. Find out more over at Neuroscientive.com.