New search engines fueled by Generative AI will compete with Google, says Semrush President

(Image credit: kalhh from Pixabay)

Semrush, together with Ahrefs and Moz, is one of the most recognized names in the world of SEO (Search Engine Optimization) and SEM (Search Engine Marketing) with its eponymous SEO tool. I interviewed its president, Eugene Levin, to find out what his thoughts are about the impact of AI on the world of search, how disruptive generative artificial intelligence (GAI) could be for an industry faced with its biggest challenge since Google launched almost 25 years ago. He talks about how he thinks Search will evolve into a multi-dimensional experience and why GAI-based search engines may be the next big thing.

1. Let's start with an easy one. What is your personal view about Generative AI as a concept?

I think this is a new technology that many people have been waiting for. For me, in terms of long term impact, this is going to be on par with personal computers, internet and smartphones. I see a lot of similarities between Generative AI and great disruptive technologies of the past.

It enables people to do things they couldn’t do before. Personal computers have democratized many areas of the economy. For example, more people could pursue software engineering that previously was an option only for those who had access to very expensive mainframes. Generative AI enables people to do creative tasks even if they lack certain technical skills. You don’t have to be a professional painter to create great concept art using Midjourney, and you don’t need to be a professional copywriter to write great blog posts using our tools.

It boosts productivity. Smartphones enabled people to be connected and improved efficiency of business communication. Generative AI can act as a co-pilot to deal with mundane tasks. For human resources professionals, it can write job descriptions and interview questions. For software developers, it can write unit tests and comments. That will free up a lot of time that can be used on more productive tasks.

It can disrupt some industries. Computers and the internet eventually displaced bank tellers. But while doing that, those technologies created more jobs. With Generative AI I can see a lot of disruption across the customer support sector. But there will be a lot of new jobs for people to train and fine tune AI.

2. What will Semrush (and its competitors) do to track what's going on in an environment where traditional SERP will disappear?

The “traditional SERP” has been gone for a while now. What we see today is many different search elements combined in a different way based on a search query and intent. Those elements have very little to do with the classic “10 blue links” interface from the dawn of the internet. Now we have pictures, product listings, structured text, videos and many other elements that are designed to provide instant answers such as featured snippets. 

I see the use of Generative AI in search as a continuation of that trend. What we’ve seen so far in Google’s closed beta is really a bigger featured snippet that is based on multiple sources instead of one. There are going to be use cases where these new Generative AI elements provide the best user experience and then use cases where new elements are not better than what Google uses today. Try searching “how old is Tom Cruise.” I doubt the search result and overall UX can get much better than what you see today.

But from a technical point of view, this is really just one more element to track. Like many of our competitors, we have been adding new elements all the time. So for us, this is just another day in the office.

Google has to rely on the authority of a page or domain, which makes backlinks increasingly more important.

On Generative AI and content

3. Someone told me that a Wikipedia page may be a good place to start to look at the future of SERP. Do you agree?

Many people predicted that once search engines start using Generative AI, informational websites such as Wikipedia would suffer the most. The idea is that if people can get an answer straight from search they don’t have to go to Wikipedia and read the article. While this idea sounds logical, it’s not what we are seeing today. For example, Bing implemented Generative AI in search several months ago. But in absolute numbers, Bing is now sending more traffic to Wikipedia than a year ago.

This is happening because asking the question and learning are two different experiences. You can’t get a full picture just by asking questions you simply don’t know what to ask. So what we are seeing now is that people spend more time asking Bing different questions and figuring out what they want to learn about. Then, once they have decided, they go to the source material to gain more extensive knowledge.

4. Generative AI is now part and parcel of the offerings of many website builders (e.g. Wix, Hostinger, Godaddy etc). What impact will that have on SEO in general?

Generative AI now being available on website builders and other platforms is great, because it levels the playing field for small and medium-sized business owners, allowing them to better compete with bigger companies that have more resources.

On the other hand those Generative AI models need guidance. They don’t know what people are searching for, what are the easy topics to rank for a particular website, and how Google sees the world. That’s why major web presence players such as Wix and Yoast partner with us o provide Keyword research capabilities to their customers. When it comes to SEO, Generative AI without data is like a brain without eyes.

Another issue is a proliferation of substantially similar content. If Jon and Jill use the same Generative AI model to cover the same topic and publish the same content, how can Google decide which article is better? In this case, Google has to rely on the authority of a page or domain, which makes backlinks increasingly more important.

5. In the notes for this interview, you mentioned new tools and new search engines that will rise as part of the GAI revolution. Can you elaborate on this?

Though there’s never been much competition for Google in the search industry, and Bing is not capturing market share substantively since they’ve operated on Open AI’s GPT-4, I believe that search could become more competitive over the next five years.

The rise of new search engines powered by Generative AI could start launching in the near-term, which is a great thing for the industry, since competition always creates elevated customer experiences. 

The advent of new search engines means marketers and business owners can show up on (and compete for) more places online, and rank highly by using best digital marketing and SEO practices, as long as they leverage the best tools on the market. We believe Google will remain the leading search engine for many years to come, but competitors may begin to take some market share.

6. How can GAI make publishers and media companies more valuable? Keen to hear your thoughts on that.

One of the key concerns of Generative AI-produced content is that the sourcing and linking to factual information within is weak, making it prone to plagiarism and misinformation. 

I believe that Generative AI is great for creating content, but it should always be human reviewed, edited, and fact checked, before it is published. Without this, we simply cannot trust the output of what is generated with AI. This will mean that today’s media publications stand to soar in value as one of the few trusted sources and destinations for factual information, so long as they don’t over-index on Generative AI. 

Over the next few years, we may see all first drafts being generated by AI, and then human reviewed. We may also start seeing media publications and websites watermark (or more clearly identify) specific sections of content that was AI produced, so that the reader can delineate between what was human and AI generated. Regardless, the top media brands should be seen as a badge of quality, trust and credibility in a now much noisier content environment.

Over the next few years, we may see all first drafts being generated by AI, and then human reviewed

On AI Content

7. Is interest in end-user GAI (e.g. front-facing Bard or ChatGPT) doomed to cool down as back-end GAI picks up (GAI-as-a-service?) What do you think?

I think that in the near term the biggest commercial successes are going to be around co-pilot capabilities of Generative AI applied across multiple industries. I expect a big splash in software development, HR, customer support and design software. 

One of the companies to monitor is, of course, Microsoft. They have unique access to OpenAI and that is their main focus now. If you watch their moves closely, you’ll come to a conclusion that their consumer product Generative AI implementations, such as Bing, are not as important for them as adding co-pilot features to virtually all of their big B2B products. Even ChatGPT, which was launched as a viral consumer product, is now heading into the direction of a business productivity personal assistant product.

8. Do you think that with its global footprint (two billion mobile devices with at least as many monthly/daily active users), its access to trillions of pictures (via Google Photos), videos (Youtube) and text (via Google Books and Google Search), Google has already won the Generative AI battle for search. If not, how do you think the threat to its hegemony will crystallize.

Google definitely has an advantage in both its ability to process large volumes of data fast and efficiently and relationships with video content creators. I think competition is feasible only in certain niches. For example, Amazon today is a search engine for products. And they use their logistics services, checkout and relationships with sellers to compete with Google in that area.

There are other markets like real estate, healthcare or online travel where some companies have unique relationships with market participants that Google doesn't have. But historically, such companies didn't have a technology that can use natural language interfaces to leverage their unique data and relationships. As you pointed out, it's not going to be a walk in the park in any case.

Desire Athow
Managing Editor, TechRadar Pro

Désiré has been musing and writing about technology during a career spanning four decades. He dabbled in website builders and web hosting when DHTML and frames were in vogue and started narrating about the impact of technology on society just before the start of the Y2K hysteria at the turn of the last millennium.