[00:02]
Claude Opus 5.5. Who wins page one when a brand new AI tool drops? Not the big sites, not the experts, it's whoever shows up first. One of my sites went shows up first. One of my sites went past 1,300 clicks in a single day using
[00:16]
this. And Claude Opus 5.5 plus my agent OS does most of the heavy lifting. Today I'll walk you through the AI SEO workflow that took a site from zero to over 1,300 clicks step by step. You'll see the five keyword types that win, the
[00:32]
search console check that finds your next page, and the mistakes that hold most people back. And near the end I'll share where I think AI search is heading. Let's get into it. First, why does this even work? When a new AI tool
[00:46]
drops, nobody owns page one yet. No big site has written the guide, no expert has an old page that's been building strength for years. So the first useful page has a real shot at ranking. That's what I mean when I say speed beats
[01:00]
authority. Now, what do you actually go after? There are five keyword types that One, the new tool name itself. People hear about something new and type its name into Google. Two, how to use plus
[01:14]
the tool name, searches want steps. Three, tool versus tool, a new tool against a competitor. Four, best keywords, best free AI model or best open source AI for example. And five, quick questions. Is it open
[01:30]
source? Is there a wait list? Is there a GitHub page? All of these are specific searches. Someone typing them usually knows exactly what they want. So step one is spotting the trend the day it drops. You can do that by hand. I prefer
[01:44]
to automate it inside my agent OS. But you don't need agent OS to start. You can do it right now in Claude. Open Claude and paste in the new tool or topic. Then ask Claude to list every searchable keyword someone would type
[01:59]
into Google about Tell it to include the tool name, how to use it, the tool versus a competitor, best tool for a job, whether it's free and open source, GitHub, and waitlist. Then ask it to rank them by likely search demand and
[02:14]
give you the top 10 to target first. I run this with Claude Opus 5.5. Anthropic released it on September 22nd as the first model in its Claude 5.5 family.
[02:26]
Claude gives you ideas, but it can't see what Google already shows for your site. For that, you need Search Console. Its performance report shows clicks, impressions, click-through rate, which is the share of people who
[02:39]
click, and average position. You can look at all of it by query. Export that keyword data as a CSV and drop it into Claude. Then ask for three things. One, keywords sitting at position 5 to 20 with over 100 impressions. Remember,
[02:55]
position is an average, so treat it as a guide. Everything I just covered is a skill you can learn, and that's what the free AI SEO course is for. It's a free course on using AI for SEO. It's for anyone who wants to learn how to bring
[03:09]
AI into their SEO work. If today's video made you want to put AI to work on your own site, this is the place to start. Learning this matters because new tools keep dropping, and people who already know how to apply AI to SEO can move
[03:23]
fast when they do. Use the link in the description to join the free AI SEO course. Now, a trending keyword only helps if people can find So here's something else that's free. Book a free strategy session. It's a
[03:37]
free AI brand visibility session. You get a custom visibility plan. You see how you show up or don't across Google, AI Overviews, ChatGPT, Perplexity, and
[03:49]
Gemini. You get a competitor gap analysis, and you get direct answers to your questions from a strategist. Think of it as a clear look at where your brand stands right now. The The is built for your brand, not a generic template.
[04:03]
It's 100% free with no obligation. The link is in the description. Book your free strategy session today. Okay, two, keywords with high impressions but a click-through rate under 1%. Google's own search console help points the same
[04:18]
way. Queries with high impressions and low clicks can show where your titles and snippets need work. Three, new keywords that appeared this week that you don't have a page for. Then ask Claude for each one to suggest a new
[04:32]
post title, a rewrite, or a section to add to an existing post. Doing that by hand works. Inside my Agent OS, it's a few clicks. I type in the topic, pick my websites, choose the date range, and run the research. Agent OS pulls in the
[04:47]
latest keywords related to that topic. Then I can deploy them to my website. Now here's the most important rule. Match the format to the search intent. If the search is a tool name, write an explainer, what it is, who made it, and
[05:02]
how to set it up. If it's a how-to, give numbered steps and show the result first. If it's tool versus tool, put the answer in the first line, then add a comparison table. And if it's a best search, give the list. This works in any
[05:16]
industry because every industry keeps coming out with something new. Here's what that looked like for me. I searched for a brand new free AI tool using a how to use it for free keyword. My pages showed up multiple times in the
[05:29]
top results. That was the how-to format on a brand new keyword. Another example is the research step inside my Agent OS. I typed in a trending tool name, picked all my websites and a date range and ran it. It pulled the latest related
[05:44]
keywords and we've built a full set of posts around one new tool this way. That brings me to clusters. Build clusters, not one-off posts. Once a keyword starts getting impressions, go after every related term around it. Take
[05:59]
the page that's getting impressions, pull every related query from search console and group them into clusters. Then decide which ones deserve their own post. One new tool can become a how-to post, a how-to use it for free post, and
[06:13]
a dozen long-tail posts. Link them to each other so they support each other. Google's own guidance on AI features also says internal links help make your content easy to find. Now, writing the post. Take the keyword
[06:27]
you found and give it to Claude Opus 5.5. Tell it your website, the keyword, and the search intent. For example, tool versus tool. Tell it to answer the search's main question in the first few sentences with the keyword in the first
[06:41]
line. Then give it the structure, section headings, a comparison table, a five-question FAQ, an author bio, and internal links. Add a case study only if
[06:53]
you have a real one. Then have it suggest a title, publish the page, and get it indexed quickly. I do this inside my agent OS, so the keyword, the draft, and the publishing all sit in one place. So why does this move fast? Speed beats
[07:08]
authority at the start. You also publish a lot of long-tail pages. Each page matches the search intent. The titles are useful, and you get pages indexed quickly. I'll be honest, going from 450 clicks a
[07:22]
day to 1,300 the next day isn't normal. I'm not promising you'll get that, but the process is repeatable. So why do people still get this wrong? Mistake one is writing one-off posts. A single post on a new tool has nothing
[07:36]
around it. Build the cluster. Mistake two is picking the wrong format. A how-to search needs steps. A versus search needs a table. If the format doesn't match the intent, the page misses. Mistake three is publishing a
[07:50]
pile of thin AI pages. Google calls it scaled content abuse when many pages are made mainly to manipulate rankings. It says that applies no matter how the pages are created. Google's guidance says using AI to help is fine as long as
[08:05]
the content meets its quality standards. So read every draft, check every fact, and make sure each page answers the question. And mistake four is never checking Search Console after you publish. The data tells you what to
[08:18]
rewrite and what to write next. Now, where's this heading? Nobody knows for sure, but here's what we can see. Google has AI Overviews and AI mode. Google's are no additional requirements to appear in them and that sites in them count
[08:34]
towards your normal Search Console traffic. I've also taken a site from zero to about 5,000 impressions inside generative AI results, not just regular Google. So here's my view. More answers will show up inside AI. Pages that
[08:49]
answer fast, stay clear, and link well will keep mattering. That's why the answer first format is worth learning now. Here's your plan. Pick one trending tool this week. Ask Claude for the keyword list. Check Search Console for
[09:03]
near ranking terms. Match the format to the intent. Write the post, then build the cluster around it. Whether you do it by hand or inside Agent OS, the steps So what do you do next? Pick one trending topic and walk through these
[09:18]
If you want a guide to learning this properly, the free AI SEO course is your next move. It's a free course on using AI for SEO, so you can learn how to apply AI to the kind of work we covered today. It's for people who want to
[09:32]
learn, practice, and get faster at it. Take what you learned today and go deeper on how AI fits into SEO. You can start learning today. Click the link in the description and join the free AI SEO course. And if
[09:46]
you'd rather have an expert look at your own brand, book a free strategy session. After a video like this, the natural next step is seeing where you stand right now. In the session, a strategist walks you through how you show up across
[09:58]
Google, AI Overviews, ChatGPT, Perplexity, and Gemini. You get a custom visibility plan, a competitor gap analysis, and clear recommendations, plus direct answers to your questions. It's the quickest way to turn what you
[10:13]
learned today into a plan for your own brand. It's 100% free with no Use the link in the description to book your free strategy session.