Andre Ribeirinho: What It Takes to Make Video That AI Can Trust

You're trained as a computer science engineer, built a social wine discovery platform over 16 years, and then pivoted into SEO and AI search visibility, founding PageRadar, right? I would say that's an unusual path into this very urgent topic that we have nowadays. What's the story? Very good question. My background has always been in, well, I studied computer science engineering, and I've always been interested in systems. And those specific systems have always been about translating complex information to help people understand that information. And sometimes I've done that with the web, sometimes I've done that with wine, that is usually a complex system for people to understand, for example, what they like. Yeah. So in 2006, I created this social network to help people choose wine. And the base has always been the same, is that how do I create a system that helps people understand, you know, a specific subject or find information? And I've always loved, for example, SEO, because it helps people understand or find the websites. I love information architecture, signals, all of these ways we deal as humans with information. Right now, doing SEO is one of the forms to, you know, it's very needed. AI is making it way more complex. And so how do we find solutions around it? Your answer actually reminds me a lot of the interview I had with Gianfranco last year. So I see a lot of similarities. Me too. You guys have a conversation right now. We were, so we met actually, just met by the bar yesterday and said, oh, you're in Japan. Oh, what is it? And then we ended up finding that we have a love for ceramics, that we have a love for, which again, had to do with systems and now the artist to do all the things. It was a very good interview from last year. So if you guys want to check it out, I would do that. I would love to watch it. I'm going back and watch it. Amazing. Your session is called Video SEO and GEO. Most people in this room understand SEO, but can you please understand what is GEO? And why does it matter specifically to video? Yes, of course. We are, you know, the world is changing. It's actually changed, of course, just recently Google changed their homepage. And it's in the process of giving access to an LLM the same way we have a ChatGPT and all the other LLMs that the way we find information is changing. Now, the thing that worked 20 years ago or even two years ago doesn't work anymore. But it still works, but it's evolving to a completely different way of searching. And I'll give you an example. That if you're now in a restaurant and you want to know which wine to pick, you can take a picture and share with Google or share with ChatGPT. And suddenly you can ask, you know, which of these wines am I going to like the most? And it's not a perfect answer, but it gives you more information than not having that sort of access. Right. But actually with video, it comes from a different perspective that having all those new systems, video has a number of unique advantages. So the kind of videos that we're talking about here are videos made by people with real experience doing something that is a product demo, a webinar. The amount of information that is in each of these videos that simply doesn't exist published anywhere is massive, right? So my argument is that if we can make those videos discoverable, if we can make them understandable, so translate them to a language that the LLMs can understand, then we will have an opportunity to unearth all this knowledge that is inside the video and hasn't been accessible because Google hasn't done a great job of showing those videos. But suddenly AI has the potential to be the best discovery channel for videos ever because you will have the content there. You will be able to search. I'm looking for a software to do this very specific thing in this industry. And if you have the video translated in a language. Right. That the LLMs can understand. Yeah. You won't be going to text. You will be going to a video of a human explaining why that's relevant. Much better. And as we were talking right now, the way that we find contents is completely different, right? So by mid 2025, nearly 70% of news related Google searches ended up without one single click to the website. That's completely different of how it was like so recently ago. What does that number mean for how companies should be thinking about their content? How companies should be thinking about video discoverability right now? It's a very good question because we live in a completely different world. And it actually has a name. It's called zero click world because there's no clicks. Yeah. We design all the systems that we built so far on getting traffic. And then with that traffic in our sites, we would be able to somehow convert in a funnel that would get us leads, new clients, you know, whatever we wanted to do with that traffic. And a big part of that traffic would come from Google. And what is happening right now is that the discovery channels is still one is still Google. The other are ChatGPT and all the other LLMs. But they are made in a way that they summarize information. Right. So they used to be that we had to do some sort of curation ourselves, go through the 10 links that Google showed. And we would go back and forth and try to understand which of the website was the one that we wanted. It doesn't happen like that anymore. We ask for types of software. It gives us 10 software that we should consider. Right. And the way these systems works by design is that they don't want you to leave until you're ready to make the conversion to buy. So we need to start thinking less about the traffic that comes to our website and more about being mentioned in those answers. Our brands, our values, our the way we work, our location of a shop that we have in a certain city needs to be showing up on those results when the target that we're trying to reach is searching for something specific. And it's a completely different way. Completely different way of doing marketing online. Yeah. And there's one one topic that is very important in this context, which is about trust. If anyone can create content saying that they are something they're not, how do LLMs and Google can say that you are the one that they should trust versus someone else? So there's a number of things that we can do from a trust perspective. One is we are ourselves and we should be saying everywhere that we have a footprint. We should be saying who we are, what we do. And it's good to be verbose about it. To say I am this person doing this to this type of target and why. A purpose is important. And you can almost go in the direction of saying you need to have a brand that people recognize that they identify themselves with and that they know that what they're looking for is what you have. And if you can do it in a way that the LLMs understand, you will go in the right direction. But it's a brave new world. Completely different than what we had before. Most companies treat SEO as an afterthought. Like adding a new thing. And then they have to go back to the title, the description and the text once the video is already made. So what are they missing out? And at what point in the production process should discoverability actually start? In this conference we've been talking a lot about strategy and doing video strategy and how much should go into the process before shooting, before picking up a camera. And video SEO and GEO should start at that level. So there's a very important word in the world of Google and the world of LLMs, which is intent. What is the intent of the video? What do we want? Of course we need to understand who we're trying to reach and all those things. But in this specific case for GEO and SEO, we want the intent of the video. Are you doing a product demo for people who are looking for, you know, like some part of the customer sales process to understand more about some sort of product? Or are you doing a support type of video? And so the way you title that video, the way you design the script of that video, the way you either invite a customer to give a testimonial versus being someone from the team doing the product demo changes. And if you can do that at the strategy level, of course, it will have much better results. You were talking about the importance of making the video visible to the LLMs, right? Because most AIs, and correct me if I'm wrong, they still cannot watch the video. Like you have to make it understandable. So what does it actually take to make a video understandable and citable by AI? It takes three things. We need to first make the video findable. We need to create the technical conditions that instead of just having a file that is lost somewhere and Google cannot find it or the LLMs cannot find it, we need to create it. It's a technical layer where we need to make sure that it's indexable, that it's findable. Okay. And it is a technical layer. The good thing is like platforms like TwentyThree, it's embedded on the platform. If you host your videos there, then it's already designed like that. So you don't need to worry about it because it's done by default. And that includes many, many different small details, technical details. But once it's well done, it's fixed. Okay. The second layer is the video needs to be understandable and it needs to have intent again. And the idea there is that it needs to be in a language, usually needs to be text instead of an audio file. Or an audio and video file. A transcription. A transcription. And there's the things that you can do there have, again, many levels. You can do a transcript. You can do key takeaways. You can do something very valuable that video has that is unique, which is questions from the audience that you had in a webinar. And you can include that in the transcript. So they are very valuable. And the idea there is to create as much text content as possible. So then you are translating the video. From a video file to something that can be read by the LLMs and Google. One of the things that, for example, is possible to do there is translations. Suddenly one video is creating content for eight, nine, ten, whatever languages that you want. And the content from all the clients that were looking for you, from all the people that were trying to reach, is now available doing that. And the third one is brand. It's trust. Again, in the world of AI where we can create whatever we want, anyone can create an article. Creating a video, yes, you can do with AI, but it's more than that. Creating a video that replaces the value of a webinar is almost impossible. It doesn't have the context. It doesn't know. It's not possible. So we need to make one. They need to be authentic so we can trust those videos. But at the same time, they need to have validation from other sources. So if the video is published somewhere, but you need to mention the video on Reddit. You need to mention the video on external review platforms if it's a software. And if you have those three layers, then the views are going to be higher. And the views are going to be visible. This is so interesting. I understand why your talk was such a success. And it's going to be available at TwentyThree on demand. And my last question for you is, if you were to say something, only one thing for someone here at the TwentyThree Summit that wants to go back to their organization and work more with video, push video forward, what would you say? If you haven't done video so far, the important thing is to start. Strategy comes usually when you know already what you want to do. Experiment with video. Do something that is shooting someone on the team and do some videos. Even internally, videos that can be shared with the internal team. But do something. Because the moment you get into it, you not only understand how to do it, but you will fall in love with doing video. André, it was a pleasure. Thank you so much. Thank you for having me.