All right, I'm going to be doing these short little voice memos from time to time on a pressing topic in digital media today. I want you to start thinking about AI SEO. Most people think SEO is dead thanks to AI, but that's the wrong way of looking at it. As someone who used to oversee $40 million in annual SEO dependent affiliate revenue in a highly competitive space, here's what I think every creator, marketer and business owner should be doing right now to adapt. First, some context. While it's true that AI is eating traditional search results, someone still needs to feed the AI. Gemini, ChatGPT and Perplexity all cite their responses, especially on transactional queries. I want to call your attention to a search for DraftKings promo code, a highly valuable term for affiliates in both chatgpt and Perplexity. Both of those AI chatbots show you the promo and bonus you can get with DraftKings, but they cite leading affiliates like Action Network, Sports Betting, Dime Rocky, Top Insider and the New York Post and generally require you to click those brands to learn more and get the code, thus giving them the affiliate attribution. So how does one feed the hungry AI? Most current best practices apply like linking, tagging and authoritative content. That's cool. Keep doing that. Let them eat. But based on hours of additional research, here's what else you must be doing for LLMs to like your content. Number one, you need to output structured data. Use JSON LD schema such as article, person or product on your content and also use same as where possible. This helps AI LLMs quickly contextualize and understand your data, improving chances of landing in snippets, AI overviews and chatbot results. If you have no idea what the heck I'm talking about, go to schema.org and learn more about the format that sort of tags webpage and document content to make it more easily ingestible not only for LLMs but also existing search engines. Number two use intent based phrases and questions. AI search is multimodal in order of appearance. That means chatbot voice and then image and video search. People ask natural language questions in chats and through voice. The more your content is framed to match these conversational queries, the better it will do. Number three Tag everything. Most content these days doesn't live on just a website anymore. It's X posts, YouTube videos, podcasts, newsletters on Beevin, Substack and so much more. There will be a gold rush for data licensing on everything from influencer content to medical records. The more everything is tagged the better. That means using alt tags on images, transcripts on podcasts, captions on videos, and descriptions on images and videos posted to social media platforms like X and LinkedIn. Hit the ad description when you upload an image to x or on LinkedIn. This is basically alt text for social media and it helps the LLMs and search engines better categorize the content you're putting out. Number four Implement LLM LLMs txt Use a tool like Mintlify or Firecrawl to generate an LLMs TXT file that LLMs can easily crawl. While this isn't industry standard yet, many expect AI bots to ingest these files much the same way search engines crawl. Robots Txt now again, if you have no idea what the heck I'm talking about. Most websites contain a file called robots Txt and they contain instructions for for search engines like Google on how they would like to be crawled so that search engine can ingest its data. Sometimes your robot may tell Google you don't want to be crawled, other times and most times you tell it you want to be crawled and how to find your stuff. LLMs txt is a proposed file format that would be more easily ingestible for AI chatbots that are crawling content. So to the extent you can include one of these on your content, there's no downside right now. Number five Be Vision and Physical World optimized if You've played with ChatGPT's latest video chat feature, then you know what I'm talking about. If you haven't, you should try it out so you can see what's coming and soon. What's the use case? Well, you look at something and then you ask questions about it. First we'll do this with phone and soon with glasses. For physical products and in person experiences, use high contrast text and easily readable fonts called optical character recognition, large 2D barcodes, QR codes, NFC and RFID tags, AR markers, whatever. Use what you can to help AI enabled devices quickly identify items in the physical world. Then make sure you connect things like QR codes on boxes to online assets and web pages that are properly tagged with schema, which we just talked about a few minutes ago. Bringing everything full circle here an example of this vision experience. Since it's so hard I think for most people to imagine today, even though technology exists, might go something like this. Someone walks down the sidewalk and looks at a pair of ski boots in a store window while wearing the Meta Ray Bans. The glasses quickly scan the QR code and present the user with the product name, price and sizes available in the store. Thanks to an integration with Shopify, the store's website that the QR code points to has product and same as schema, so the glasses are able to present the user with similar products and reviews. All of this is possible because the glasses easily identified the product they were looking at. This is what's coming in Search. It's not just about online anymore. The more we walk around with real time AI cameras, whether through our phones or soon ubiquitously with glare, you need to the AI needs to be able to look and quantify that data quickly. All right, so to recap, what do you need to do? 1. Structure your data 2. Use natural language phrasing 3. Tag everything. If there's an option to add a description or a Tag, do it. 4. Implement the LLMs TXT file so AI chatbots can better crawl your content and then five implementations be thinking about physical search, especially if you're a store owner or you do something in the physical world. All of this is just scratching the surface. It's very early days, but I want you to know search isn't dead. You just have to think about it differently.