Dr. Noah St John explains how to maintain an LLM-powered AI assistant yourself. Do four things weekly: audit responses for accuracy, rewrite web pages into Q&A format, create bullet-point answers with statistics, and use structured headers. Skip any of these and your assistant loses visibility.
All right, so let's look at the DIY path because some of you are going to be watching this video, listening to my voice, and you're going to say, this all sounds great, Noah, I'll just go and do it myself. Cool. All right, that's fine if you decide to do it that way. But here is the challenge, right? If you decide to do it yourself, the essential weekly tasks you must commit to for DIY success: audit assistance across buyer prompts. So you need to regularly evaluate the performance and accuracy of assistance the LLMs, responding to diverse buyer inquiries every week. You have to keep track of this. Remember, I just talked about freshness, right, and ubiquity. Then you need to rewrite web pages into a Q&A format. You need to convert existing web content into clear question and answer structures to improve usability and searchability. Then you need to build bullet answers with stats and citations. You need to create concise bullet point answers supported by relevant statistics and credible citations for authority. And finally, structured headers for scanning. Or you need to organize headers strategically to enhance readability and allow users to quickly scan content. All right, so you are creating this content and you have to do it every week. If you stop, the AI assistants will move on and you will go back to being invisible. Okay, this isn't a scare tactic, this is reality. So here, let's look at the pain of doing it yourself because it is a pain. There's no way around this.
