Less than a month ago, Google I/O shook up the world of SEO. One of the most important announcements was Google Search Generative Experience (SGE). Google SGE is Google’s new vision for search experience, with a specific focus on organizing and summarizing what Google considers to be the best result into a concise response.

With SGE, users are able to:

• Ask detailed questions and get detailed response, which is a big shift away from keyword-oriented search.
• Summarize complex topics into an overview and get helpful links to read further
• Ask follow-up questions in a chat experience
• Turn searches into creative activities such as new document drafts and more

If you haven’t seen or read it yet, be sure to check out Google’s white paper on SGE.

It covers the following topics in details:

1. What is SGE
2. How SGE works
3. Applying generative AI responsibly
4. Known limitations
5. The future of search and user participation

At 18 pages, it’s a quick read and will be worth referring back to as the technology evolves.

SGE is still in beta and it’s not fully available to all users yet. This will likely be at least a 6 month or longer rollout (just guessing).

We believe that Google is using this period as a testing and development phase, not just for its LLM technology but, arguably more importantly, to understand the impact of LLM-based chat features within their core search product.

In particular, right now everybody is focused on simple chat interactions but where the real power of this integration is leading is what we are seeing in the open source world with projects like Langchain, AutoGPT, BabyAGI, and others is something far more transformative: agents.

Google rightly understands that search is the initial action of so many behaviors and agents provide the ability to intelligently automate many human behaviors and workflows that currently begin — and end — with a search interaction. The rest of the action is completed on sites to which Google sends that traffic.

We all know that Google is interested in keeping users on their properties longer but this doesn’t mean that they will be able to completely monopolize all possible workflows and automation.

This is where I think search — and SEO — over the long term are heading. Companies wishing to play in Google’s ecosystem will need to position themselves as producers of value via APIs and their own agent-focused plugins providing the data, workflow options, and tools needed for Google to select them as providers in their own agent-driven workflows.

The “ten blue links” as well as presence in SERP features will continue to play a role for a few more years but this is where things are heading and companies doing business on the web need to get a lot more sophisticated very quickly.

Marie Haynes had a great post last week covering what many of the influential SEOs are saying about their own experiences with SGE.

Some of the observations that were made:

• Google isn’t necessarily pulling in the top results for use within its AI-generated responses
• The responses generally include websites for further reading and these, also, are not necessarily (and often not) top ranking results
• The websites that are included in the responses tend to be UGC-heavy and / or informational
• This will have an impact on search terms with commercial or transactional intent, making it simultaneously more important to understand search intent and also expect lower CTRs (maybe) for these terms
• It all depends on how many users actually request the auto-generated responses vs. skipping it and going for links the way they do today – this is the real experiment
• Local results will change in a big way
• Lack of attribution appears to be an issue

At the end of May, Google provided clarification on how topic authority is used for news sites. They noted that this is not a new ranking system and, in fact, it has been in use for several years already.

In the vendor world of SEO, SaaS companies (such as ourselves) involved in SERP monitoring are looking at, generally, how AI is going to shape the SERP and, specifically, how we will be able to monitor the presence and inclusion in Google’s SGE within the SERPs.

Right now, there isn’t a good way to do it at scale.

The primary reason for this because Google is requiring a signed-in account and enabling the SGE features from Search Labs. This will make it easy for them to shut down anyone indexing their SERPs (at least for these purposes).

Customers paying for SERP data likely won’t want to pay the extra fees associated with maintaining large numbers of spoof accounts to get these results.

So, right now, there isn’t much you should expect to see in tools (other than browser extensions).

How much does it really matter? Right now? Probably not very much. As we’ve seen with Google product roll-outs in the past, if it really matters, there will usually be a way to get the data. Google might make it generally available down the road so it could be possible to get SGE results without being part of Labs.

And if they don’t, then it will still be a level-playing field.

This is one of the reasons we’ve spent so much time aggregating a variety of data sources, including Google Search Console, into one data view, because there is never one perfect source of information in the SEO world.

Finally, as we discuss in the next section, SGE is currently available for English-only results.

As we mentioned above, SGE is still in beta and it is early days. It appears for English results only, we even tested this on some Japanese keywords in our SGE-enabled account and no SGE results appeared:

“range rover” in Japanese:

If you haven’t already and, if you have the option in your country, go to Search Labs and sign up for the beta. You’ll need to enable it for your account. You’ll need Chrome or the Google app too.

Then just start playing around with it.

What I found personally is, I don’t find myself using the AI-generated results very much yet. I am a regular user of ChatGPT for specific types of searches, mostly related to code-generation and summarization of complex topics.

Also, as mentioned in the beginning, if you haven’t read Google’s whitepaper, it’s worth reading.

Make sure you’re subscribed to our blog as we will continue to explore the growing of AI on search and many other areas of our lives online.

You can follow us on Twitter and connect with me on LinkedIn.