SearchGPT vs. Google: What It Means for SEO

What happens to search once SearchGPT is part of it? DigiAssets’ full review

The SEO world is shaking ahead of the upcoming SearchGPT launch.

Can you already picture a world where search becomes a conversation?
Instead of scrolling through dozens of results, we could just ask “What’s the best way to lower my electricity bill?” and get a personalized answer based on comprehensive, up-to-date information. That’s the future SearchGPT promises.

But what do the prototype and usage data from 10,000 users so far tell us?
Does it look better than Google?

Is this a golden opportunity for small sites to stand out?
Is this the end of “keyword research”?
And how will we do SEO in an era where SearchGPT might become the main search engine?

Data and takeaways from SearchGPT usage so far

SearchGPT’s search results are 73% similar to Bing’s results
SearchGPT’s search results are only 46% similar to Google’s results
26% of the domains ranking in SearchGPT get no traffic from Google
About 32% of all SearchGPT results come from media sources
For media-related queries, more than 75% of results come from media sources
Less than 1% of searches returned inaccurate results
We can infer that SearchGPT favors established, authoritative sites
Backlinks are a significant ranking factor in SearchGPT.

The surprising advantages

Information is summarized and condensed very well for users: SearchGPT does a good job of searching, researching and gathering information from different sources, then presenting it in a very clear, concise way.
That means if you search for “effects of climate change,” you’ll likely get a comprehensive summary covering scientific data, economic impacts and possible solutions, all in one place, instead of scrolling and sorting through search results yourself.
An opportunity for small sites: 26% of SearchGPT results come from sites that don’t rank high on Google. Think of it this way: a small travel blog about Italy can show up alongside major tourism sites when someone looks for information about Rome. That’s a significant advantage.

The significant challenges

Less diversity in information and sources: While Google shows dozens of varied results, SearchGPT tends to show a limited number of sources. Take a search like “ketogenic diet.” Google shows recipes, scientific articles and forums, while SearchGPT may focus only on the definition plus a few health benefits.
Information manipulation: As we already know, the biggest challenge with AI is FACT CHECKING and keeping it from “hallucinating” text and information. When search becomes a dialogue between a person and an AI, there will always be room for the AI to interpret, phrase and rewrite things, which can lead to inaccurate information.

What do I expect to happen to SEO with SearchGPT in the picture?

It’s an opportunity and a challenge at the same time.

On one hand, there’s a real opportunity here for small sites with quality content to stand out.
Search isn’t going to die anytime soon. Quite the opposite: the more users can get results that are more accurate, more niche and better matched to their search intent, the more global search volume will keep growing. That convinces me the SEO industry is only going to get more sophisticated, on several levels, in the future.

Also, remember that AI can’t invent information. Without human content, there’s no AI that can base data on it, summarize it and deliver it conveniently. I’m convinced that for exactly this reason, as long as AI search engines want to stay in the game, they depend on content creators and site owners, and they’ll keep giving them the credit they deserve and partnering with them on publishing.

On the other hand,
if you haven’t started preparing strategies built for AI search, now is the time.

My two cents on SEO strategies for AI search engines

More focus on semantic optimization: It looks like there will be a big focus on understanding the intent behind searches rather than exact-match keywords, just like what happens today when you send prompts to ChatGPT and it compensates for your mistakes. Using related words and concepts will help AI understand the broader context of our content.
AI needs to read your intent even when you don’t search or ask the question in the most precise way. So there’s no telling whether, in the long run, we’ll keep doing “keyword research” the classic way we’re used to, or whether pulling search volume and KD data will even be relevant. But we’ll definitely need to get creative when researching the topics we want to write about.
Will comprehensive, in-depth content get more credit?: My hypothesis is that AI will prioritize comprehensive sources. One long, in-depth article on a topic may be more effective than several short ones. Remember that AI can consume long, comprehensive content and condense it into a single sentence, so cutting corners on content won’t be a good idea.
Clear content structure: We’ll need to focus on a logical, clear content structure, with precise headings and subheadings, to help AI summarize the information efficiently.
The E.E.A.T concept will win: Just like with Google, there’s no way this engine will crawl every source on the internet to find its answer. There will be a clear preference for sites that invest in their brand and have experience, trustworthiness and authority behind them, signaling to the AI that this is a site to pull information from. We’ll need to invest more in building a professional profile of supporting assets for the site and in quality links.

In the end, it seems SearchGPT will only highlight and sharpen what we’ve always known: quality, relevant, useful content. Bottom line: more accurate results for the user.

Still, despite SearchGPT’s promising potential, OpenAI faces serious challenges on the way to becoming a real Google competitor. There’s a long road ahead in improving information accuracy, expanding the engine’s crawling and data collection capabilities, improving the query system, and of course, it wouldn’t hurt to build a continuous learning mechanism based on user feedback.

SearchGPT vs Google