agentos.guide › blog
How To Do GEO For A Website: A 9-Step Quick Start You Can Run From Your Agent OS
Here's how to do GEO for website owners in nine steps, using the free Hermes workflow I show in my latest video.
GEO means generative engine optimisation, which is making your business the name that Google AI, Gemini or ChatGPT gives when someone asks a question.
The short version is that you export your Search Console data, let Hermes find the topics you can win, write case-study pages with an SEO skill, then fact-check and publish.
This quick start shows each step and which part of it an Agent OS can run for you.
If you'd like help first, book a free one-to-one GEO, AEO AI audit and get a custom game plan.
How to do GEO for a website: the 9 steps
This is the whole method on one screen, with the part an Agent OS can take off your plate.
| Step | What you do | Who does it in an Agent OS |
|---|---|---|
| 1 | Log who the AI answers name for your buyer questions. | An agent runs the questions and you read the log. |
| 2 | Export your Google Search Console performance data to CSV. | You click export once a month. |
| 3 | Ask Hermes to turn the CSV into GEO keyword and topic ideas. | Hermes builds a report artifact in the workspace. |
| 4 | Write case-study-led pages with an SEO skill. | Hermes drafts from your real results using the skill. |
| 5 | Publish different angles on several sites and platforms. | Hermes drafts each angle and you approve each one. |
| 6 | Publish on new topics within 24 hours. | An agent watches your niche and flags new topics. |
| 7 | Check crawler access and basic schema. | An agent runs the checklist and you fix what it finds. |
| 8 | Fact-check every line by hand, then publish. | You do this yourself, every time. |
| 9 | Measure clicks and AI impressions, then repeat. | An agent compares this month with last month. |
Steps two to four are the Hermes workflow from my video.
Steps four to six are the three moves I call the answer engine, which are be the source, be everywhere and be first.
What GEO is and why it needs a workflow
GEO stands for generative engine optimisation.
It's the work of making your business the name that Google AI, Gemini or ChatGPT gives when someone asks a question.
Old SEO gave ten blue links, and position seven still got some clicks.
AI answers give a short list of maybe one to three names, so there's no room for number seven.
The AI reads the pages that rank and sums up what they say, like five friends recommending the same restaurant.
That's why GEO is a repeatable workflow rather than a one-off trick, because you need several good pages to agree on you over time.
Google's own guide to optimising for generative AI features says that, from its point of view, this is still SEO.
The free stack I use
In my video I run this from my Agent OS, but you can run the same process with Hermes Agent on its own.
| Piece | What it is | Cost as its own page describes it |
|---|---|---|
| Hermes Agent | An open-source agent from Nous Research that does the analysis, drafting and publishing. | Open source under the MIT licence. |
| OpenCode | An open-source AI coding agent that I use as Hermes's brain. | It says free models are included. |
| Google Search Console | Google's tool for your site's search performance, including the CSV export. | Google calls it a free service. |
| Your SEO skill | A checklist the agent follows on every article, written from a case-study angle. | Mine is roughly 100 points and lives inside the AI Profit Boardroom. |
| Your case studies | Whatever you work on day to day, such as tests, client jobs and experiments. | They cost you the work you're already doing. |
When I recorded the video, DeepSeek V4.1 Flash was free inside OpenCode, but free models change, so check the list on the day you set it up.
🔥 Want the exact AI SEO setup I run inside my Agent OS?
Inside the AI Profit Boardroom you get the installable Agent OS, step-by-step video tutorials, four coaching calls a week and 3,400+ members building real automations.
Quick start: steps 1 to 3 in one afternoon
Step 1: log the baseline
Write ten to twenty questions a real buyer would ask about your business.
Run them in Google AI Mode, look for an AI Overview, and ask ChatGPT and Gemini as well.
Save the names, the cited pages and whether you appear into shared memory, so every agent can see the starting point.
Step 2: export Search Console
Open the Performance report, pick your date range and use the export button to download a CSV.
Export the queries and the pages, because the agent needs both.
This is Google organic data, so it won't show every AI tool, but it's the best free proxy for what's already working.
Step 3: hand the CSV to Hermes
Drop the CSV into the workspace and send Hermes this prompt.
Can you analyse this and, based on what's actually working, give me keyword and topic ideas to rank inside AI search engines?
In my run, Hermes built a report artifact showing total clicks, the brand share of clicks, the non-brand click-through rate and my September clicks.
It pulled out what was working, such as comparison articles, my Agent OS clusters, the free AI community topic and my Antigravity and Hermes topics.
It finished with averages and a verdict on my best pages, which is where your first GEO topics come from.
📞 Want help reading your topic report?
Book a free one-to-one GEO, AEO AI audit and my team will look at your site, your competitors and the keywords you're closest to winning.
Steps 4 to 6: the answer engine inside an Agent OS
Step 4: be the source with an SEO skill
I give Hermes my SEO skill and access to the sites, such as WordPress or Netlify, so it can draft and publish.
My demo prompt in the video is this simple.
Create an SEO article using your SEO skill and the keyword "GEO examples" and publish to aiprofitboardroom.com.
The skill writes everything from a case-study angle, and that angle is the whole point.
AI Mode describes me as widely recognised as a top practitioner because of public testing and deployment, which is on my screen.
To be named as the expert, you have to show that you've actually used the thing you're writing about.
If you run experiments in Claude, you can export them to Hermes, or write the article inside Claude instead.
Step 5: be everywhere
I run multiple websites, each with a different angle, all pointing to the same answer.
That's how I can show up two to four times on page one for one search, and my subreddit, AI SEO Insider, ranks on page one too.
An agent can draft each angle quickly, but every version must be genuinely useful on its own.
Google's guide warns that separate pages for every search variation, made mainly to manipulate rankings or AI answers, break its scaled content abuse policy.
Step 6: be first
I try to publish within 24 hours of a topic appearing.
When Jev AI dropped, I covered "Jev architecture" and "Jev waitlist" that same weekend, and my Jev AI guide explains the tool itself.
This is the easiest job to hand to an agent, because it's just watching and flagging every morning.
🔥 Want the exact AI SEO setup I run inside my Agent OS?
Inside the AI Profit Boardroom you get the installable Agent OS, step-by-step video tutorials, four coaching calls a week and 3,400+ members building real automations.
Steps 7 to 9: checks, fact-check and measurement
Step 7: crawler access and schema
Google says there are no additional requirements to appear in AI Overviews or AI Mode, but the page must be indexed and eligible for a snippet.
Have an agent check that robots.txt doesn't block Googlebot, and that OAI-SearchBot and PerplexityBot are allowed if you want ChatGPT search and Perplexity.
OpenAI's bots page says sites opted out of OAI-SearchBot won't be shown in ChatGPT search answers.
Perplexity's bots guide says PerplexityBot surfaces and links sites in its search results.
Check the Search Console Search generative AI control as well, which includes your site by default.
Google's guide says no special schema or llms.txt file is needed for its AI features, although normal structured data still helps with rich results.
Step 8: the human fact-check
Google updated its page on using generative AI content on 1 October 2026.
It says it's "critical to manually factcheck and review all AI-generated content" before publishing, including titles, descriptions, structured data and alt text.
Google explains that generative models predict likely words rather than retrieving facts, which is why they can get things wrong.
So in my Agent OS the agent drafts, but a human signs off every number, name and meta field before anything goes live.
I'm adding a manual fact-checking list inside the AI Profit Boardroom for exactly this step.
Step 9: measure and repeat
Re-run the baseline questions every couple of weeks and compare them with the log in shared memory.
Search Console now has a Generative AI performance report showing impressions from Google's AI features, and Google says it rolled out to all sites as of 31 August 2026.
AI Mode does send traffic, and I show it on screen in the video.
One of my sites went from zero to more than 1,300 clicks a day, and two other sites are on the same trajectory.
Those are my results from my own screens, not a forecast for yours.
The weekly GEO loop I'd schedule
Once the first pages are live, the job becomes a loop that agents can run on a timer.
| When | Agent job | Your job |
|---|---|---|
| Every morning | Scan launches and news in your niche and flag anything new. | Pick one topic worth a same-day page. |
| Every week | Draft one case-study page and one different-angle version for another platform. | Supply the real result and fact-check both drafts. |
| Every fortnight | Re-run the baseline questions and log who gets named. | Read what changed and decide the next topic. |
| Every month | Read the new Search Console CSV and refresh the topic report. | Export the CSV and approve the new topic list. |
The fact-check column never moves to the agent side, however good the model gets.
📞 Want a strategist to pick your first GEO topics?
Claim a free AEO SEO strategy session and leave with a custom game plan built around the keywords you're closest to winning.
What I'd skip
Google's guide has a mythbusting section, and I agree with most of it.
It says Google Search ignores llms.txt files, so I don't spend time on them for Google.
It says you don't need to chop content into tiny chunks or rewrite it in a special style for AI.
It also says chasing inauthentic mentions isn't as helpful as it might seem.
Nobody can guarantee what an AI answer will say, so I'd also skip anyone who promises a placement.
What's inside the AI Profit Boardroom for this
If you want the system rather than the idea, the AI Profit Boardroom has it.
You get my Agent OS, my SEO skill and checklist, an AI SEO automation section and four coaching calls a week.
There's a member map with 3,400+ members, and you can DM me inside, where I can even take a look at your website.
If you'd rather talk it through first, book a free strategy session and see the keywords you're closest to winning.
Where to go next on agentos.guide
My GEO examples guide shows four real AI answers that name me or my agency.
My free GEO audit guide helps you check your site before step one.
My Claude AI SEO rank #1 guide covers the Claude version of this workflow.
Also On Our Network
FAQ
What is the quickest way to start doing GEO for a website?
Log who the AI answers name for your buyer questions, export your Search Console data to CSV, and ask Hermes to turn it into keyword and topic ideas for AI search. That takes about an afternoon.
Can I do GEO for free?
Yes. Search Console is a free Google service, Hermes Agent is open source under the MIT licence and OpenCode includes free models, although which models are free can change.
Do I need an Agent OS to do GEO?
No. I run it from my Agent OS, but the same process works with Hermes Agent on its own.
What should an agent never do in a GEO workflow?
It should never publish without a human fact-check, invent case studies or results, or copy the same page across several sites. Google's 1 October 2026 guidance says AI content must be manually fact-checked before publishing.
Do I need llms.txt or special schema for GEO?
Not for Google. Its guide says Google Search ignores llms.txt files and that there is no special schema for AI features, although normal structured data still helps with rich results.
Is the free strategy session a written GEO report?
No. It is a one-to-one call where my team looks at your site, your competitors and the keywords you're closest to winning, then gives you a custom game plan for Google and AI search.
Final word on how to do GEO for a website
Export your Search Console data today, run steps one to three this afternoon and schedule the weekly loop.
When you want a second pair of eyes, book a free AEO SEO strategy session and get a plan to rank in Google and AI search.
That's how to do GEO for website growth from your Agent OS.
About Julian
I'm Julian Goldie, an AI entrepreneur, SEO expert and the founder of the AI Profit Boardroom, which has 3,400+ members.
I help business owners scale with AI agents, automation and SEO.
My YouTube channel has 400,000+ subscribers, and I run Goldie Agency, a seven-figure SEO agency.
Related guides on agentos.guide
Get the whole Agent OS
The installable system behind every guide on this site — dashboard, agents, memory and pipelines — updated as new tools land, with 3,400+ members and four coaching calls a week.
Join the AI Profit Boardroom →📺 Video notes + links to the tools 👉
🎥 Learn how I make these videos 👉