SEO Implications of Search Engines Morphing Into AI Chatbots

While the big AI companies are delving into legal gray areas trying to find new sources of data for their Large Language Models, the situation for website operators and publishers is much more black and white. Traffic to websites is drying up — and a full-on drought could be about to hit, as search engines morph into generative AI chatbots.

To find out more about how search engines are changing in the AI era, and what website operators can do to adapt, I spoke to Jim Yu, the founder and executive chair of BrightEdge, a corporate SEO platform that has been running for seventeen years.

Are Perplexity’s Citations Now Driving Traffic?

BrightEdge recently released an interesting study of Perplexity, an AI chatbot platform that is trying to compete with Google on search (when I spoke to Perplexity CEO Aravind Srinivas back in January, he admitted he was “a big Larry Page fan” and had modeled his search engine on PageRank).

What immediately stood out to me in the study was the contention that people are increasingly clicking Perplexity’s citation links. “BrightEdge’s analysis shows that referrals from Perplexity to brand sites are growing at nearly 40% month over month since January,” stated the media release. “This is the first evidence that Perplexity is not used solely for generative content purposes, but also as a search engine that refers users to sites through citations.” (emphasis mine)

This struck a chord with me, as a regular user of Perplexity (disclosure: Perplexity gifted me one year of its Pro service while I was researching the January article). When I first started using Perplexity Pro, the citation links were visible in the chatbot replies to questions I posed, but there was little prompting to click on them. But recently, I began noticing that Perplexity is prompting users more to click those citation links. In many of my replies now, I get phrases like “according to the search results” and “the key evidence is” — which tend to make me want to click the accompanying links more.

I asked Yu if this might be why referral traffic from Perplexity to websites is growing? While he wouldn’t be drawn on that particular question, he did say that Perplexity has staked a clear position in search that is the polar opposite of ChatGPT.

“I think at the heart of the Perplexity AI experience is that it is providing a well formulated, well researched answer to your question,” he said, noting that it’s much more transparent than ChatGPT in that regard. He added that Perplexity gives you the sources, but it also provides “a nice AI summary of it.”

Clearly, the main value a user gets from Perplexity is that summarization of the source material. But it could also be in the company’s interests to get people to click the citation links, as a way to help the user gain trust in Perplexity. In any case, the results from BrightEdge on referral traffic is encouraging from a publisher perspective.

The Giant Awakens: Google’s Search Generative Experiences

While Perplexity has captured the interest of many early adopters in tech, the 800-pound gorilla of search has been quietly preparing its own AI-based search. Google’s “Search Generative Experiences” (SGE) has been an optional feature in Google Labs for nearly a year now, although for most of that time it’s been geo-restricted to US users. While it has been expanded to other countries recently, the UK (where I live) has yet to see it. So I asked Yu for his impressions of Google’s SGE?

“Google’s actually somewhere between the two,” he said, referring to ChatGPT and Perplexity. He thinks SGE is “much more transparent” than ChatGPT. He added that while “it’s more closer to Perplexity than it is to the GPT model,” it’s “less prominent” in how it includes citation links.

Yu has seen Google go through a series of iterations to refine SGE over the past eleven or so months.

“In the early days, the experience wasn’t cached,” he noted. “So it was pretty slow when it loaded. Now it’s cached, it’s pretty fast. We also see they’re tuning it, industry by industry.”

Notably, the industry tuning includes implementing “guardrails” in industries like health and finances, as a form of risk management in how it implements AI search for those industries.

“The number one place that they use AI is your health,” said Yu, regarding Google’s testing of SGE. “The reason for that is they’re very careful about that. So if you search for a disease in the AI version [of Google Search], it puts a disclaimer right at the top — hey, you should consult a doctor. The second thing it does is it takes over a lot of the page and it clearly weighs certain sites like the CDC (the Center for Disease Control) and NIH (the National Institute of Health). They [also] weigh Mayo Clinic, major health institutions.”

While those same sources are also ranked highly in Google’s current, non-AI, search, Yu’s point is that they’re even more prominent in the AI search results. He says that this extra weighting, alongside the disclaimers, are because “the AI generates certain opinions” and so Google naturally wants to protect itself.

This seems to imply that Google SGE will place even more emphasis on authoritative websites than ever before. Which means it’ll be harder for new entrants — let’s say, a new health information website — to rank highly in Google’s AI search. So I asked Yu what web operators can do to optimize for AI search, given that it’s clearly going to be harder to get cited (especially in industries like healthcare).

“If you’re doing good content and you’re a trusted brand, you have a good foundation to build for,” he said. “The future means that you’ve got to get much better at understanding what conversations do you as a brand have a right to win in? Make sure you’re creating unique perspectives on that — the heritage of your brand, where you have authority.”

Will Publishers Benefit from AI Citations?

It’s all very well being cited in search results, but of course, websites want traffic to flow back to them from those citations. While BrightEdge’s study has found an increase in referral traffic from Perplexity, that won’t necessarily happen with Google SGE — especially since Yu said that SGE isn’t displaying citations as prominently as Perplexity. So I asked him what website operators — and particularly media businesses, which rely on that traffic — can do to adapt to the AI search era.

He replied that media sites “have to get very sharp about what is the topic area where you are going to have the right to win in, where you have the most expertise.” So brand authority will become even more important, because “the AI engines have to attribute trust somehow to different sources of information.”

He said that in order for AI engines to organize sources of information, they will increasingly rely on “authority graphs” that build metadata about publishers. While this already happens with traditional search engines, it’s now being fed into the LLMs — “ all the stuff you want to surround the content with, so the AI knows how to do the attribution,” said Yu.

All of this means that being a citation in Google SGE or Perplexity (or Bing’s OpenAI-based solution) is increasingly going to drive your search optimization strategy. Indeed, Perplexity’s Aravind Srinivas said a similar thing back in our January interview. According to Srinivas, the more that Perplexity’s product cites a certain web page, the more important it becomes.

“The next generation of rankings is going to be citation,” said Yu, “and you want to be, as a publisher, the original cited source for information, because everything else later is going to be generated from the AI around the original source. And so the graph will be about being the original source on different topics. So I think that’s really important for publishers.”

Yu’s point is that being the original source — the “trusted source” — is going to be vital for publishers going forward. Unfortunately, there’s no guarantee that traffic will flow back to them even if they are that original source for certain topics. But there is at least a little hope that Perplexity is beginning to drive traffic back to publishers, and we can only hope (and pray) that Google follows suit.