James Dooley: how AI recommends businesses in 2026. Nowadays, artificial intelligence and the LLMs are recommending lots of different companies, whether the Fortune 500 companies or solopreneurs and tradesmen. So Jason Barnard, you started talking about answer engine optimization back in 2017 with your white paper with Trust Pilot and webinar series with Semrush. Everyone thought you was crazy, including myself. You even coined the term answer engine optimization. Now it's mainstream. What do you see today that others still don't? Jason Barnard: I think what I'm seeing today that others don't is that answer engine optimization was started in 2017 in in a real sense in that the machines the Google started giving answers. A lot of people talk about generative engine optimization today but actually it's AI assistive engine optimization. generative is what they do in terms of what how they create those answers. But we're talking about how does the AI assist people and then the future is going to be AI assistive agent optimization. So I'm not looking at um at just today uh AI AI assistive engine optimization. And I'm looking at the future AI assistive agent optimization. And the that's the point that people are missing is that GEO generative optimization which is the super popular term right now in my opinion is already out of date. James Dooley: Yeah, that makes Um, and then missing the fact that we're going to move very quickly through AI assistive engine optimization to AI agent AI assistive agent optimization. James Dooley: Yeah. So, with regards to recommendations, I want to just break things down for anyone who might be watching this that might not be very advanced with LLMs or AI and stuff like that. There's one or two questions that I've got which is like is being recommended by AI whether that's chat GBT or GE Gemini or Claude or Publexity is that the same as ranking in Google or is it completely different? Jason Barnard: It's significantly different. Um with ranking Google you're thinking about pages in AI and AI overviews and chat GPT and Google AI mode. You're thinking of passages. Google call them passages. as Microsoft call them chunks. It's passages, chunks of information that the AI is pulling out. So you need to break your pages down in your own mind into these chunks. So in order to get into those results in the AI, you need to be thinking about the different parts of the page that answer specific questions within the overall question. People talk about fury query fan out. I talk about cascading queries. And it's basically the machine asks itself what are the related queries around this that will help me answer the query better. So if I ask who is Jason Barnard the cascading queries the theory the query fan out is going to be who is Jason Barnard Jason Barnard reviews Jason Barnard credentials Jason Barnard career so that it can answer the question better and the key here is uh a friend of mine was talking about the difference between uh a child saying something from memory and a child having to look something up in an incite opedia. If you can get the LLM to answer without looking at the search results and without looking in the knowledge graph because it's able to say it from its existing knowledge, you're winning the game because you're going to be the preferred solution because it prefers answering off the top of its head. James Dooley: Yeah, for sure. On there with regards to cascading queries, um I think there's a Google patent for query network which is part of query augmentation um which has been around. Everyone's become obsessed with query fan out as if it's a brand new feature and query augmentation has been around for a long time um within Google. But I've got a follow on question then with regards to two different companies. They're both trying to do the passages of content and the chunks and they're both trying to rank for specific queries. What makes AI trust one company and one website over other website more? Jason Barnard: It's a it's a lovely question because there are multiple layers going on here. If if the AI doesn't have good grip on either of them, let's say we've just got two. If it doesn't really understand either of them, it will just look at listicles and pick the ones on the listicles because it doesn't have better resources. And that's an easy way to win the game short term if the situation is that the competitor is not understood. But if they understand both of them, i.e. It knows off the top of its head which one of them whi what what each one of them does and how they answer the question or solve the problem. They're going to look for the one that has the most credibility signals that it has picked up. So it's e and we call it a cali cube ne a t because we add notability and transparency to experience uh expertise authoritiveness and trustworthiness. James Dooley: And when we're looking at artificial intelligence recommending you, how important is it with the information that's on your own website, also known as a firstparty source, versus what others are saying about you on a third party source? How important is both of them? And what's is is one more important than the other? Jason Barnard: Well, the third party sources don't mean anything if there isn't a first party source. So, you have to have the first party source. So that's non-negotiable. Then you've said what you've got to say and you've said this is who I am. This is who I serve and this is why I'm the best. The machines won't believe you on your own good word. You need the corroboration. So both are necessary. Without the first party, the third party means nothing. And without the third party, the first party means nothing. Start with first party, build third party. There's no point in repeating 100,000 times on your own website exactly the same information. You need corroboration on third party sites. James Dooley: Yeah. And then another thing is you talk about being found versus being recommended by the LLMs. What's the difference? Jason Barnard: Well, you can be found by the machines as they crawl around the web. Super important. You're in the index. You're in Google's index. in Bing's indexed or you're you're being used by track GPT and the crawlers coming to your website, they found you. Do they actually respect you? Do they care about you? Are they and we were talking about this earlier on, are they your well-trained employees? The answer is probably no. And the difference there is training. If you train them to use the information they found about you in the way that your real employees would be using it, you're going to win the game. train the AI to use your information in the way that you would train your employees to do it. James Dooley: Yeah. I mean, here in episode number two in this playlist, we speak about the digital salesforce AI employees recommending you at that 11th hour is the difference between you winning and losing a job to a competitor. We took in one of the other episodes we spoke about brand entity SEO for tradesmen and local people. And this leads me on to the next question with regards to a small business like a tradesman versus the big brands, the Fortune 500 companies. How can little small business try to compete against big brands with AI recommendations? Jason Barnard: Yeah, I love this just because throughout the history of the internet and I started in 1998, so it's not the whole history of the internet that I've been through. I missed out on a big chunk at the beginning. But you know, I've seen these opportunities where people suddenly say, "Oh, the little guy can compete against the big guy." And it happened in 98 when I started. Um, it's happened multiple times as algorithms update. And the AI revolution started out with people saying that and now people are saying, "Oh, actually famous companies or big companies are always going to have the advantage." The trick has always been niche down. If you can niche down, you can always beat the big guy. Niche down, prove that you're the best for that very specific niche, and once you've mastered that, expand out. But trying to beat the big guy toe-to-toe is a losing battle. James Dooley: Yeah. So if someone watching this now um whether they're a soloreneur, tradesman, high netw worth individual, a business owner, whoever it is that's watching it, how what is a key takeaway now if someone wants to start being recommended by the LLMs, whether it's chat GBT, Perplexity, Claude, Gemini, what's step number one? Jason Barnard: If I were if I were just starting now, here's a nice trick. Go to your go to the AI you use every day. It knows you. It knows who you are. It knows what your business is. And it understands what you're trying to achieve. Do a quick brainstorm with it. Ask it. Describe in couple of paragraphs. Who am I? What do I do? Who do I serve? Why am I the best? Get the answer to that and then ask and then sorry and then ask it. If Jason Barnard were to advise me on a strategy for my company or myself, given that, what would he say? He could do it with any of them. This is the really sweet part. All of them know my methodologies back to front, upside down, round and round. They will give you a very good answer as to what I would advise. That's my advice. James Dooley: Great bit of advice. Anyone watching this, this is episode number nine, an 11p part playlist series with regards to how AI recommends you and obviously advancing on how to try to get recommended by the art force intelligence. Jason Barnard, it's been an absolute pleasure. Jason Barnard: Thank you, man.