Introduction

My name is Tom Mason, and I'm the founder of Awareness AI, where we help organizations understand how they are represented and recommended in different models of AI.

What Is GEO (Generative Engine Optimization)?

So this is now being called GEO, Generative Engine Optimization, and it's being seen as the evolution of SEO. So you might have heard different terms for GEO, you might have heard AEO, LLMSEO, but it seems to have settled now at GEO.

and in this live demo today I'm basically going to help you understand how to get a baseline and a benchmark of your organization standard in geo so I'm going to run a live prompt which is going to target recommendation and

representation and I'm going to explain to you how to which prompts to use how to analyze them and then how to take first steps to improve the question

Why Organizations Should Care About AI Search

really is is why should you care how your organization shows up in an AI AI model such as ChatGPT and Gemini.

And the data really backs this up because 50 % of consumers now are using AI -powered search.

So this means that people are asking ChatGPT and Gemini questions such as, what organization should I choose? How would you describe X organization? And this is therefore leading to AI becoming a first point of contact and a first point point of sale for organizations.

It's shaping an organization's reputation before a user even ever gets to their website. So it's becoming really crucial to understand what AI is actually, how AI is representing your company and how it's recommending your company.

And another key figure is 750 billion in revenue will be impacted by 2028. And this is from McKinsey 2025.

From SEO to GEO: How the Search Journey Is Changing

So when looking at the change from SEO to geo, it's interesting to look at the change in consumer search journey so the traditional search journey for consumers was often a question in a search engine

such as Google a user will then search through multiple links they'll do the research and they'll make their choice on what to choose but now with AI

powered search they are asking a question or prompt in a large language model such as chat GPT they're gaining an instant AI summary and then they're making the choice.

So what's changed? So the things that have been taken out really are searching through multiple links or websites and research.

These two areas here have been combined into an AI summary and AI now, a large language now, is now doing the research and searching through multiple links for users.

So this

session today is really going to focus on the AI summary part of the new search journey and how to understand it and improve it.

A Practical 3-Step Workflow: Prompt, Analyze, Improve

So it's going to be a three -step workflow of prompt analyze and improve so the initial prompt is

going to target recommendation and representation so you know if you asked top organizations in your industry this is going to instantly show you if you are being recommended or how would you describe your organization this is going

to tag representation and then we're going to analyze the prompt to see if to help you understand if you are being recommended or represented accurately and then we're going to go through areas of how you can improve and this will be

looking really at sources and understanding where the information is being inferred from.

Live Demo Setup: Using Temporary Chats to Reduce Bias

So you're just going to go now on to ChatGPT and to give a initial overview we will be using a temporary chat and the importance of

using a temporary chat in this scenario is that it hasn't got any past saved memory so if you were to ask about your organization in your chat in your usual chat it will have it will have passive memory about your organization which will then create bias and influence the result so using a temporary chat stay

Recommendation Prompt Example (Italian Restaurants in Leeds)

important and I'm going to start off with a recommendation prompt of I'm

looking to eat out in Leeds tonight please can you recommend me an Italian restaurant so this is going to show basically ChatGPT recommending me an an Italian restaurant in Leeds.

So you instantly see here that it has recommended me one Italian restaurant, which is Zucco in Meanwood. And then it has given me two good backups.

And now, there are definitely more than three Italian restaurants in Leeds, that is known. But AI has only suggested me one confidently and two backups.

So now this is really showing the narrowing down of choice for consumers, consumers, because very few people now would go into Google and search for more Italian restaurants and look for different choices.

They would probably treat this as a trusted source and choose one of the three. So it's showing why it's becoming extremely important for, in this case, Italian restaurants to understand if they are being recommended and

understand how they can improve this. So that was the prompt and the initial analysis. So

How to Analyze AI Outputs: Sources and Trust Signals

Sources are really important because the sources is how AI is inferring trust and how AI is creating its output. This is the information it's using to formulate its response.

So if you were an Italian restaurant in Leeds and you were analyzing this output, you're looking at first -party sources and third -party sources.

So first -party sources are things you can control, such as your organization's website. and then third -party sources are sources that are shaping your reputation but you don't have direct control of.

So first -party sources here you can see which are being cited are Zucco and Live in Italy which are two websites and then you have third -party sources such as TripAdvisor and then if we scroll down a bit more here you can see we have OpenTable and other different ones such such as Facebook down here.

So when you're running this prompt for your organization, you want to be looking at which sources are being pulled from, and you want to be understanding where the potential opportunities are for you to be getting in within these sources.

So for example, if you're an Italian restaurant, you would want to be targeting Facebook, and you would want to be targeting TripAdvisor.

Why 'Top 10' Listicles Influence AI Recommendations

And recommendations in AI Interactivity and GEO are often very dependent on something called list calls. calls.

So this is often top 10 best Italian restaurants in Leeds, or if it was in your organization, it would be like top 10 cybersecurity firms in Leeds.

AI often uses these list of calls, PR list of calls, to create recommendations.

Testing Representation: Follow-Up Prompts and Brand Perception

So now we're going to ask a follow -up prompt, which is going to target representation. So we're just going to put this prompt in here, basically asking why did you choose the top recommended company so here it is

forming a judgment about Zucco and it is basically representing it it's saying it has it's saying it's a family Italian restaurant it is intended for several dishes and it's got strong external validation but what's interesting as

well is that it's going to provide a risk assessment of Zucco before Zucco has even had the chance to provide its first interpretation it's going to talk about the main drawbacks and now when you follow up with this prompt for your

organization you want to be looking at the drawbacks and you want to be seeing what AI is saying negatively if you like about your organization so if we look here it doesn't seem too bad there seems to be very little drawbacks but for

For example, it's less convenient if you stay near the station, or often small plates. So this is really being perceived by AIs as a negative thing.

And now if we go into the sources, so if we analyze and how to improve, you see in the number one source is the website, which is very good. You see in the second source is the third -party source, which is the Guardian.

The Outsized Impact of Reviews and Third-Party Mentions

But what I find mainly about working in geo and representation presentation is that third -party sources are often reviews. That is often how AI is shaping its perception and reputation of you. And often AI sees absolutely everything.

This is really important to understand.

So before with SEO and the traditional search journey, if you had one negative thing said about you on the 10th page of Google, most people wouldn't see it. And a business would know that. Customers aren't going to see that.

But now an AI sees absolutely everything across the whole web. So if you have one negative thing said about you on the 10th page of Google,

it's going to bring it up in chat and it's going to tell you about it if you've got a thousand positive reviews and one negative review that is unmanaged that is unmanaged then AI is going to see it and it's going to pull it so understanding your whole presence across the whole web and understanding where there could be little gaps in your reputation is really important to understand

and I think what's interesting about this whole example with Zucco is that all of us have now just gained a first interpretation of Zucco and I'm not sure how many of you or how many of you have been or how many of you know it but Zucco probably isn't

the best Italian restaurant in Leeds it might be a small Italian restaurant in Leeds which has amazing service amazing food that could be miles better than Zucco but they have no online visibility and they maybe they don't have a website so they are missing an opportunity here and it will affect really the how successful the business is because they don't understand their their geo presence so just to

Key Takeaways and Next Steps

wrap it all up now um i think it's important to go over that ai is a tool and it's not a verdict and ai can produce hallucinations and misinformation and that is a big part of what we're looking at so it's important to understand that ai can be wrong it results can vary and it's

always important to send to check it out but whenever you're working with a large language model such as Chachapiti or Gemini.

Conclusion

And that is everything from me. Thank you very much for listening.