SEO in the LLM Era: Threats and Opportunities

As Google and Microsoft continue to push AI search into their product, there seems to be a common (and not all that crazy tbh) fear that AI search will destroy the internet by cannibalizing traffic with "stolen" content. However, I think that view is exaggerated and lacks context of some of the things Web development, design, and SEO pros can do to adapt.

What kinds of AI Search Products Exist and Why could they be Problematic for Site Owners?

Right now, I think there are two major AI search experiences worth noting, Google's Search Generative Experience (SGE) and Microsoft's Copilot (which I will call 'Bing/Copilot' in this article so it's clear I'm not talking about Github Copilot). The brand names and experiments might change over time, but these two examples highlight different types of LLM-juiced search experience that'll likely be relevant for at least a while.

SGE is an experiment that uses AI to generate an answer to whatever you Googled (in applicable situations) and puts it at the top of your Google search results page. In a lot of ways, this is more "passive" than some other LLM "experiences" you may come across online, but also very in-your-face when turned on because it shows up at the top of the page.

Bing/Copilot is a GPT-based AI chatbot that has access to Bing Search results. It provides a powerful ChatGPT-like experience with the ability to intelligently do internet searches based on user intent, location data, device metadata, and other personalized sources to get context. Bing's tiny search engine market share might might make this seem irrelevant at first, but Microsoft also seems very intent on baking it into Windows 11 whether people want it or not.

With both of these types of "AI Search" products there is an obvious risk that people will not click through to high-quality websites if the AI generated results are good enough.

This is a Cyclical Issue, not a Doomsday

Did Google kill all sports websites when it started integrating live sports scores directly into the their results pages? Of course not, and yet Google has almost certainly sucked away some portion of traffic that would have clicked through the various sites in the past.

For decades, long preceding "modern AI", Google has spent a lot of engineering resources making its search results more useful, making countless little "widgets" (or whatever word you want to use) that appear at the top of search results. Even though LLMs are a whole new can of worms from a technology perspective, the fact remains that Google is a search and ads company that has faced this basic problem before. In fact, for companies willing to spend some cash, this opens up an exciting new ad market to reach consumers (Bing/Copilot has already had paid ads for a while now).

Local SEO will be Extremely Powerful

I came across this article by digital marketing agency Global Reach when researching for this, and I think it offered some great advice.

Expanding on their suggestions further, you should attempt to use good seasonal keywords as well in your website to help AI search pick out your business in relevant situations.

I'll share an example of this that impressed me from my personal experience:

A couple days ago, I asked Bing/Copilot for help trying to plan a couple more skiing days before the end of the season (I wanted to see if it would plan a trip to Killington or South America for me). It told me that all of the ski resorts near me in New Hampshire were already closed for the season, and it searched "Summer activities near Thornton, NH" for me. Local businesses with the right keywords got a free shot to convince me to click through to their site with an AI customized pitch.

For businesses in the tourism and service industries, it will be essential to exploit this combination of local SEO and seasonal keywords. It's basically like a free salesperson!

Write for an Audience that includes Generative AI Models

Broadly speaking, I think many of us will find this different compared to writing for people and search engines directly. I don't have a list of do's and don't to provide you here... I think we're all still learning on the fly here largely (feel free to share your experiences in the comments). But now, your webpages are potentially a way to insert your ads into other people's AI prompts. Don't ignore the power of this opportunity.

I think the proliferation of chat AI will drive the continued popularity of "landing page" website designs that push as much meaningful content about the business as possible to a single homepage. While separating content into different pages might be more optimal from the perspectives of UX and traditional SEO, an AI search might only look at one page in your site when generating an answer for a user.

Social Media Engagement and Open Graph

If clicks from Google directly are going to be tougher to earn, I think that obviously increases the value of social media. I see 2 areas to focus on that could be impactful for a lot of companies

1. Leverage social media platforms to algorithmically build a following/community that trusts you and who may eventually check out your website. Tik Tok and Youtube Shorts both seem to be really good platforms for this. LinkedIn can be good too. Let these platforms' algorithms and recommender models find an audience for your niche content
2. Use Open Graph protocol to provide social networks like Facebook and X/Twitter/whatever with metadata about your pages that helps fit them into the social graphs. If you're using a CMS for your website, you can likely find a SEO plugin that also supports managing Open Graph tags

These points are probably not new to a lot of readers, but their importance will only continue to grow as we need to look for Platforms outside traditional search engines

Also, keep an eye out for industry-specific search engines and directory websites. There are a lot of platforms like this that let you make a profile so people can find your business on the sites they use to find specialized services.

Lastly, business owners and SEO pros should remember that Youtube and Tik Tok are major search engines for a lot of people, making it essential to play the game I describe in point no. 1 above.

Develop Rich, Interactive Web Experiences

Naturally, simple text content is pretty easy for language models to "steal." If you invest in building rich, dynamic content, like interactive Javascript infographics, for example, you can give people a reason to actually come to your site that AI can't steal on the fly.

Conclusions

Although the integration of Large Language Models into search engines poses threats to individual website owners and the web platform as a whole, these challenges are not insurmountable. Just as the web has evolved in response to past innovations, so too can it adapt to this new era of AI-driven search.

Web development, design, and SEO professionals must stay ahead by embracing strategies that highlight the unique value of their content. Focusing on local SEO, creating content with generative AI in mind, leveraging social media engagement, and developing rich, interactive web experiences are all vital steps. By doing so, businesses can not only mitigate the risks posed by AI search but also find new opportunities to connect with their audiences in more meaningful and impactful ways.

Do you have any strategies you're using to exploit AI search for your business? Please feel free to share your tips and experiences in the comments! I think this area will evolve quickly, and I am curious to hear more real-world stories!

Thank you for reading!