For the last year, people engaged in search engine optimization have fallen into two categories. There are the ones who believe that basic SEO fundamentals still win the race to the top of both organic search engine and AI results, and those who insist that new rules have to be followed in tandem with some of the old ones to optimize for AI prompts and Google’s AI summaries that show on the top of search engine results. People in this category refer to optimizing for AI as generative engine optimization, or GEO.
While everyone in the second category doesn’t do this, some agencies and people claim traditional SEO is completely dead and only they possess the magic beans to get you to the top of AI search for a healthy retainer. The truth actually lies in the middle, but if you’ve been in the traditional SEO camp, Google has just handed you a trophy.
Google recently released a definitive guide to optimizing for generative AI features on Google Search, which seems like it was written to torpedo the sales pitches of the “magic beans” crowd and assert that SEO fundamentals still win the day in generative AI search in its own index. It is a goldmine document in that Google gives us more of a roadmap than it ever has for how to rank in its search index and in Gemini.
Where the AEO/GEO/SEO Confusion Comes From
ChatGPT, Gemini, Claude, and similar AI apps draw from various resources to create answers, images, proposals, or whatever you choose to create with them. One of these resources, and perhaps the largest in play, is Google’s search engine index.
The GEO-first crowd doesn’t completely write off Google’s index as a source for AI apps, but usually claims that content has to be written differently for AI by using FAQ-based formatting and/or “chunking”, where data is broken up into small bites to make it easier for AI to digest. Google has asserted in its guide that this isn’t the case, at least when it comes to its own products.
Non-Slop Content is Still Important
Google points to its own guide for SEO best practices for content and structure, and it's much simpler than the other stuff you’ll find elsewhere on this topic. It includes:
“Create helpful, reliable, people-first content.”
This is probably the most hotly debated topic in SEO and GEO. Good, helpful content takes time. It usually takes a person, at least for editing. Nobody loves the amount of time and effort it takes to even just edit and fact-check an AI-generated first draft. However, according to Google, it absolutely has to be done to rank.
You can read more about what Google recommends for content creation in its new guide here. In short: don’t mass-generate AI slop, edit AI-generated items, create unique content that stands out from the crowd, and make sure it helps your readers in some way.
If you own a small business and don’t have a lot of time, Gemini does a great job of creating first drafts for most topics. Give yourself time to edit these drafts. A massive protip for you is that many media and academic sites actively block Gemini and other AI tools from accessing their sites, but a Google search will turn them up. Just do a quick Google search for your topic and pick an academic or media source or two to punch up the “helpful content” meter on your creations.
There will be facts there that AI doesn’t know about, and since roughly half of the content on the web is now AI-generated, it will make your take unique - something Google also recommends in its new guide.
It’s also worth noting that since the dawn of the internet, good content has consistently been a major ranking factor. That has never changed, and Google has explicitly stated that it’s not likely to.
The Technical SEO Generative AI Does - and Doesn’t Need
Technical SEO is the term used to describe back-end code items that are done to improve your site’s rankings. Generally, a web developer will do these tasks if asked, they’re generally not a part of your web dev’s job function unless you require it. Some things you can do yourself, but they’ll be faster at it.
The technicals you do need
1) Clear technical structure
Your web pages should, at the most basic level, have technically correct elements in place and follow coding best practices. This is where a content management system (CMS) is handy, because using one generally means that you don’t have to worry about this.
2) Maximize your crawl budget
Everyone who has a website has seen “Discovered - Page Not Indexed” in their Google Search Console. If you don’t know what that sentence means, hook up Google Search Console to your website and crawl your pages, then you will.
This crawl guide gives you a list of technical to-do’s if you see that common error, and Google explicitly recommends following them in its guide.
3) Good user experience (UX)
UX is up there with content in terms of being a major ranking factor for Google since nearly day one of its inception. It has dovetailed off into being its own specialty, and is important not just for search engines, but more importantly for conversions and sales. Following Google’s tips for optimizing it helps with nearly every aspect of your digital presence, not just search rank.
The technicals you don’t need
1) LLMS.txt
Google specifically calls out LLMS.txt files and other markup meant to speak directly to AI code. They don’t work for its code specifically. Some have maintained that they do work for other AI resources, but until AI tools say they need them, probably best to pitch this make-work project out.
2) Overly structured schema
One of the big items that the AI-first search crowd includes in a sales pitch is schema.org markup, which means additional detailed markup in your site’s code that tells a machine what each section contains, as well as some other information. If your site has zero schema, it’s still a good idea to have a very basic schema in place, at least on your top-level pages.
You can get the details for how to do it here. You can ignore anything in the advanced section, except maybe on your home page.
Google is basically saying in its new guide that it can tell what is on your site without having schema in place, but still recommends basic schema as a best practice.
The real death knell for the “hours of coding your schema” salespeople wasn’t in Google’s new guide, but a recent study by Ahrefs, a popular SEO tool, which asserted that schema didn’t move the needle for AI citations in 1,885 tracked web pages. However, having basic schema is still great for accessibility, so if you have invested in it, it wasn’t entirely a wash.
With that, we’ve given you and your web developer enough homework. There’s absolutely no reason not to bring in outside help if you need it. One of the things Google’s guide says in closing is to prioritize effective SEO strategies over "AEO/GEO hacks." This may be a bit heavy-handed on Google’s part, but they’re just trying to save business owners from getting hosed.
It is true that most SEO agencies have rebranded as AEO/GEO agencies, mostly to attract business since that’s what people are looking for - not in a bid to misdirect. They’re (hopefully) still using the same old traditional SEO practices that Google recommends, while keeping an eye on anything new that might be on the horizon.
If you want to optimize your operations alongside your AI and SEO search relevance, we’ve got you covered! UpCurve Cloud is a trusted Google Partner to some of America’s best-known businesses, government agencies, and nonprofits. Contact us today to migrate to Google Workspace and get the best AI tools for your business in the process.
Contact Us to Learn More about Transforming Your Business