How To Rank in Google's AI Mode (with Examples)

Watch on YouTube

Category: Digital Marketing

Tags: AI modeContent marketingDigital PRSearch optimizationSEO strategies

Entities: Exposure NinjaGeminiGoogleSemrushThe OrdinaryZugu

Building WordCloud ...

Summary

Transcript

00:00

In May, Google finally released AI mode and the world went ballistic. At least the people who actually know about this.

We've been talking about AI mode on the Exposure Ninja channel for months now and preparing our clients for its launch. But now the rest of the world is starting to realize that actually search

00:16

is changing profoundly now. Given that Google tends to migrate features from its AI tests into main search, this is coming.

features from AI mode are going to find their way into main search and be impacting marketers this year. So, of course, if you're anything like me and

00:32

the ninjas at Exposure Ninja, your immediate thought goes to, "Oh, okay. How can we get this thing to recommend our business's products and services?" Well, that's exactly what we do for clients.

And we're going to share some of our early observations on how to do exactly that today. To do this, let's think about the difference between AI mode and regular search.

Because there

00:49

is some discussion that in order to get ranked or recommended in AI mode, all you need to do is do regular SEO, but that's not quite true. In regular search results, you have web pages ranked by an algorithm.

It uses loads of different criteria like relevance, authority of the website, links, experience of the

01:07

person or entity writing the content. Now, AI search tools like AI mode also have a web search component.

They will go and run web searches in the background. In many cases, they will run multiple different web searches.

So, yes, SEO is important to make sure

01:23

you've got visibility in those web searches. But tools like AI mode have an additional layer of complexity, and that is the AI bit.

And AI mode uses a version of Google's Gemini large language model. Gemini has its own knowledge which it can contribute to the

01:39

answers. It also reads and analyzes the content being ranked in web search.

Now, this might seem like a small or obvious thing, but actually it's fundamental to understanding AI mode and how to get it to recommend your business products or services. As we discussed in our

01:54

previous video, which was an introduction to AI mode, here are the things that we're recommending our SEO and content teams at Exposure Ninja focus on in order to improve our clients visibility inside AI mode. Some of these things are just good SEO.

So, we're

02:10

going to focus on some of the things that are new to this world of AI search ranking. These are going to be the battlegrounds of the future and where brands will be built or destroyed in the coming years.

We're going to start right at the end here with brand mentions, links, and PR targeting key positioning.

02:27

Brand mentions are, in my opinion, the biggest lever as we move into this new world of AI search. I'm not alone in that observation.

This data published by AHS found that branded web mentions also followed by other brand factors had the strongest correlation with visibility in

02:44

Google's AI overviews. And this is where we come back to this idea that AI search optimization is not the same as regular SEO because of this AI bit.

Now all of these AI search tools are powered by large language models. Let's think about what a large language model is and how

03:00

it works. Fundamentally, these tools like chat GBT are prediction engines.

By learning patterns, they predict sequences of words and concepts. Let's look at an example.

If you go into chat GPT and type in have a break, have a chat GBT completes the sentence Kit Kat.

03:17

Well, how did it do that? Well, the large language model behind chat GPT has been fed vast amounts of information on the internet.

And when it's ingesting all this information, it's linking up different concepts, different groups of words, different patterns. These are stored in what are called embeddings.

03:34

Now, imagine the large language model as a huge brain. And inside this huge brain, you've got different concepts all dotted around.

Now, concepts that are frequently seen next to each other in the training data have stronger links inside this brain. So, have a break and KitKat are very closely connected inside

03:52

this brain. But ChatGpt also knows that this is a tagline because have a break, have a Kit Kat and tagline are also concepts that are very closely related.

And ChatGpt didn't need to go online to run a search to find this information. It already had it in its training data.

04:09

Okay, Tim, who cares what has this got to do with my brand? Well, let me show you another brand example.

And you may have seen me use this one before. If I ask chat GBT for the best value skincare brands, it doesn't need to go online in order to recommend our client, The Ordinary.

It knows The Ordinary's

04:25

positioning and also their key products. How does it know this?

Well, these concepts of best value skincare brands, The Ordinary, are linked very closely in the large language model's brain because it's ingested so much information about

04:40

these three concepts and seeing them so closely linked together. And you'll notice that often you get consistent results whichever AI tool that you're using because they've essentially been trained on a similar amount of data.

So here in AI mode, if you ask the best value skincare brands, you also get the

04:56

ordinary. And it's precisely this type of pattern recognition that we need to exploit if we want to get these tools recommending our brands as frequently as possible.

In fact, to the extent that some of these other brands are actually being evaluated according to their similarity to the ordinary. So again,

05:11

these concepts of best value, skincare brands, the ordinary are so closely linked. And by the way, you don't need to be a huge brand in order to do this.

Exposure Ninja is a digital marketing agency that offers a free website and marketing review carried out by a real human. And we've talked about that for

05:28

years and years. So there's lots of content online linking those concepts.

So when you ask AI mode, which digital marketing agencies offer a free website and digital marketing review carried out by a real person, it spits out exposure ninja as does chat GPT. Look at that

05:44

reputation. Okay, so how do you engineer this for your business or service?

Well, there are really two components to this. The first is to be super clear about the concepts that you actually want to link to your brand, product, or service.

Now, that's very simple for me to say, but it can be a very difficult challenge for a

06:00

lot of companies to figure out. Usually there are two ways of doing this.

You can either link yourself to a particular type of audience, for example, car insurance for young female drivers. Alternatively, you might want to be known for having a particular position in the market.

Maybe you're lowest cost

06:16

or highest spec or you have great technical experience or your products are used by celebrities. And if you're struggling to work out which concepts you're known for or want to be known for, a really useful exercise can be just to talk to your customers.

We do this for our clients at Exposure Ninja

06:31

through something called a brand accelerator where we actually call up our clients customers and ask them what are they known for? Why did you choose them over their competitors?

And what we tend to find is that consistent threads keep coming through. Another important exercise can be finding out what these AI tools already know about you because

06:47

maybe you can just accelerate into some of those sentiments. I love using the Seamrush AI tools for exactly this purpose.

And full disclosure, Semrush let me use this tool free of charge if I show it in the videos and I'd be using it anyway because it's my favorite tool. But here you can see an analysis of vanguard.com and we can see that on

07:03

Google's Gemini AI tool sentiment about Vanguard is overwhelmingly positive because of their industry-leading low expense ratios i.e. low costs calculators and visuals.

So if we're Vanguard, we might be pushing further

07:18

into that space, talking frequently about how we are the lowest cost. And Samrush allows you to see how your visibility compares on different AI tools.

And by the way, you can get a free trial of Seamrush at thankyouninjas.com. Check it out.

It's amazing. If you don't like it, you can cancel before you have to pay anything.

07:34

So nobrainer. So once you're clear on your positioning and what you want your product, service, or brands to be known for, the next step is to make sure that that positioning is found across the internet.

Because as these AI tools are ingesting all of that information, the more times they're seeing these concepts

07:50

linked in that content, the increased likelihood you have of being spat out when a user asks for best value skincare brands. I mean, obviously, if that's what you're going for.

And how do you do that? Through digital PR.

Here's an example of one of our clients. If you ask AI mode for the best iPad Air case,

08:08

it recommends the Zugu case. Why does it recommend this?

Well, firstly, Zugu, iPad Air case, and best are concepts that are closely linked in the AI tools brain because they are found close to each other lots of times online. Secondly, we've helped Zugu get lots of

08:25

visibility in the types of underlying web searches for this query. For example, getting them featured in this ZDNet article, which positions them as the best iPad Air case.

And you can see that this is one of the sources for these search results as well as Zugu's own website content. Getting featured

08:42

through this type of digital PR is not easy and it's not cheap. Either if you're doing it yourself or you're working with an agency like Exposure Ninja to do it.

It takes time, energy, and relentlessness. So, I'm not going to tell you that you can do this in half an hour every other Sunday.

It's just not

08:59

going to happen. This requires a consistent and concerted effort over months and in some cases years.

I'm just being honest. But if you want this next generation of search to be consistently recommending your brand products and services, this is what you need to do.

09:15

Of course, getting talked about on Reddit core userenerated content sites also help. But there's no getting around the fact that the more often you're seen online, the more closely these concepts that you're targeting are linked together, the higher your likelihood of being recommended by these AI tools.

09:31

Don't get me wrong, this doesn't mean that you don't have to be some huge business in order to be getting recommended by AI mode. That is absolutely not the case.

And we do have some clients which are smaller and aren't able to go for huge digital PR campaigns. How they can fight and win in

09:46

this world is by going more niche by focusing on certain product verticals. And one of the ways that they can do this is through content.

Your content strategy must change for AI mode. Gone of the days of producing very thin generic top offunnel content.

What these

10:02

AI tools love is high expertise, high authoritiveness, high experience content that targets the middle and the bottom of the funnel. Let me give you an example.

If I ask AI mode what kelp lawn fertilizer I could use, I see our client

10:18

golf course lawn here being cited as a source for AI mode. And that gives Ron the owner a chance to influence these answers.

Now if we have a look at the article that has been used as a source here, we will see is exactly the sort of thing that all of these AI tools love to

10:35

site. Why is that?

This is a really detailed article. This is way beyond the sort of generic content, the generic information that these AI tools can come up with themselves.

This is clearly written by someone who knows this topic in detail and has firstirhand experience

10:52

of it. In fact, we can even see that because not only does Ron recommend particular products in this article, but there's also a video at the end showing him actually implementing this stuff on a lawn.

His bio at the end of the article talks about the fact that he's got a certificate in sports turf grass

11:07

management from the University of Georgia. He is a true expert in this space and that makes this content something that the AI tools are not only more likely to recommend and site, but it also makes it more likely that a user would click on this article to get something beyond what the AI tools can

11:24

give them. Now, our content writers work with Ron to produce this.

So, this doesn't mean that top expert in your business needs to produce all of the content themselves. Our content writers work with Ron, but it's that real expertise that needs to shine through.

Like I say, gone of the days of just producing generic content. And if you're

11:40

just asking Chat GBT to write all your content for you, unfortunately, generic is exactly what you're going to get. Okay, what are the topics that we need to actually write about then?

Well, as we covered in our previous video, AI mode uses something called the query fan out technique where when you ask a

11:56

simple question to AI mode, it breaks that question down into lots of different background searches, it runs those web searches and then compiles all of the information gathered from those into a single answer. This means if we want to increase our chance of being recommended, we need to make sure that

12:13

we've got content which targets as many of those underlying searches as we possibly can. In AI mode, that's difficult to do because we don't actually get to see what those background searches are that are being carried out.

But then we've got the Dale hack because if you run that same search

12:28

in Google's Gemini AI tool, it does in fact show you the background searches that it's running. That doesn't mean that AI mode is using exactly the same background searches, but it's a pretty good indication of the types of background query that are going to be searched.

Your job then of course is to

12:44

produce detailed rich content about each of these different topics and get that content ranking in web search. That way when AI mode does its query fan out technique, it keeps seeing your website, your content, your expertise again and again and again and the chances increase

13:01

of it recommending your brand, your products or your service when it gives that answer. Right now you might be feeling a little bit overwhelmed about AI mode and this new future of search.

It seems like such a departure from SEO and it is. Whilst good SEO is the sort

13:17

of foundation of AI mode visibility or chat GBT visibility, really there is a new approach required here. But don't get disheartened.

This same disruption is happening to every business in every industry. There are established businesses that will be left behind because they are so used to getting

13:33

loads of organic traffic from Google. They don't adapt their strategy to this new type of AI based search.

and there are new businesses that will pop up and steal market share as a result. The key, of course, like it is for any disruption, is to embrace it and work out how you can adapt your strategy to

13:49

get in front of this trend. This should be caught in your marketing effort.

The cat is out of the bag here. Google is only going in one direction.

Search is only going in one direction. Search behavior is changing.

It already has changed. And therefore, whilst ignoring this stuff might seem like an option

14:05

now, in the coming years, it really won't be. Of course, feel free to contact the team at Exposure Ninja if you want any help with this.

This is exactly what we do. This is our main focus as an agency is helping our clients win in this new world.

Just go to our website at

14:20

exposurinja.com and drop us a note if you're interested in working with us. If you've got any questions at all about AI mode, drop them in the comments and we'll do our best to answer.

Until next time, see you