Nathan Gotch

How I create SEO content (that ranks) with ChatGPT:

🤖 Prompt 1:

Create a table that does the following for the content below:

- Extract the top 50 keywords (include 1, 2, and 3-word phrases)
- Give them a relevance score based on this core idea: “[KEYWORD]“
- Include how many times the keyword phrase is mentioned (frequency)

[COMPETITOR CONTENT]

⎯

*Repeat this process for the top 3 competitors to get an average.

🤖 Prompt 2:

⎯

Create an article that's between [# OF WORDS] words for the keyword phrase: “[KEYWORD]“ using the details below:

1. Use the keywords below to craft a unique and optimized article (between [# OF WORDS] words)

[NLP KEYWORDS]

2. Use the heading ideas below, but make them unique

[COMPETITOR HEADINGS]

⎯

*Use a free word counter or ChatGPT to get the average word count.

⎯

Final steps:

1. This gives you a nice base to work with. Now you can inject your own unique experience and expertise into it.

2. Don't forget to edit (we aim for 95 + with Grammarly)

3. Optimize with your favorite on-page SEO tool

4. Repeat to build more topic authority

5. Acquire relevant, high-quality backlinks.

If this was helpful, then:

➤ Follow me because I have more helpful SEO content coming soon.

➤ Like and share this so others can learn too 👍

I just read "GEO-ready content."

Please stop it.

Pages using standard SEO practices work perfectly fine to influence AI generated responses and citations.

Do this:

⊕ Make sure your page is easily crawlable, blazing fast, and HTML

⊕ Place your keyword phrase in URL, title, H1, first sentence, and a variant in the first H2

⊕ Cover the most relevant entities (topics) using a tool like Rankability

⊕ Don't worry about word count. Focus on topic coverage.

⊕ Add a TL;DR and use many descriptive headings

⊕ Inject relevant internal links

There is no magic GEO dust to sprinkle on your pages.

80% of "GEO" (driving brand visibility in AI chat platform) is just doing SEO well.

I'm launching a new podcast called The AI Search Report and I want to interview the most elite AI SEO experts on the planet.

Who do you want to see on our podcast?

There's one commonality I see among the most successful SEO practitioners:

they just simply love the game.

When I do SEO, I feel like I'm playing a video game.

Time becomes irrelevant.

Or,

Maybe it's just me 😅

11 ways AEO is different than SEO:

1. Tracking - tracking brand mentions in answers, not blue link rankings or local pack rankings. Also, tracking citations used during retrieval (see point 6).

Note: you should track traditional SERP rankings for the sole reason that search is the biggest source of retrieval.

2. It’s very hard to win with just your website. AI platforms use many sources for training static corpus + RAG to gather consensus. Your website is just one source.

3. AI answers are non-deterministic which causes answer and citation volatility. Compared to SEO / traditional search engines, which don’t change as often.

Same with the local pack, which stays pretty static for extended periods of time.

4. Extreme personalization - there is personalization when you use Google, but ChatGPT is next level.

5. Prompts - long-tail phrases are not new, but the length of queries is dramatically longer and has significantly more context. Especially when voice is used.

6. Reverse engineering citations - Competitor analysis is not new, but reverse engineering sources of retrieval is unique because it isn’t always extracting from search.

7. Retrieval not exclusive to search engines - Doubling down on the point above. Retrieval can occur through document uploads, API, direct retrieval to websites, etc.

8. Optimizing for brand visibility in answers (not clicks) - KPIs are more indirect like branded searches.

9. Local - doing well in Google’s local pack doesn’t necessarily translate to ChatGPT visibility.

You didn’t need reviews or consensus to rank in Google SERPs, but you need it to be recommended as a brand in AI answers.

10. Query fan-out - attacking every angle of topic isn’t new, but the underlying tech is different compared to how traditional search works.

11. Demand data is invisible - No, there's no such thing as "prompt volume." No one knows true demand data for specific queries in ChatGPT.

All you can do (right now) is use search data as a proxy for estimated demand.

Good luck out there!