What I Learned About SEO and AI Search in 2025

Posted on November 18, 2025 by Marketing Stuff

Recently I was part of a group asked to give a presentation to our department about changes in search engine optimization, specifically about declining traffic from Google and the rise of conversational AI tools like ChatGPT. This is something I’ve been following for quite a while, and while gathering research for the presentation it occurred to me that this may be helpful for other folks too.

If you’ve been working in SEO and following all these changes like I have, this info probably won’t be new for you. But if you’re a marketer and you’re curious about how Google AI overviews and ChatGPT are changing SEO, here’s a breakdown of what I’ve found.

In my job at Columbia Southern University, most of my focus has been on driving organic traffic to our website through blog articles. We’ve done this through traditional SEO tactics like:

  • Finding keywords with high search volume and low competition.
  • Matching users’ search intent with quality content using EEAT (expertise, experience, authority, trustworthiness).
  • Earning backlinks.
  • Technical SEO tactics like meta description optimization, proper headings and tags, etc.

For the most part, these tactics have worked. We reached peak traffic levels in 2023, but then the numbers started going down.

Changes for Google

Google launched AI overviews as an opt-in feature in 2023, then to all users in 2024. Since then, informational content that summarizes what’s already widely available online can now largely being replaced by these Google AI overviews.

Google now sends less traffic to websites due to “zero-click searches.” Users frequently get their answers through AI overviews and move on.

  • Search queries that show AI overviews result in a 34.5% reduction in clicks (Ahrefs). We’ve seen similar percentage decreases in Google traffic for our blog content.
  • Search queries that don’t show AI overviews also result in less clicks, hinting at broader changes in search behavior (Seer Interactive).

Even with the rise of other tools like ChatGPT, Google use isn’t declining. It’s growing (SparkToro). It’s just not resulting in as many clicks to websites. Google is increasingly becoming the place you go AFTER you find out about a brand on social media, YouTube, podcasts, email newsletters, broadcast news, advertising, conferences, etc. And yes, ChatGPT.

The Rise of Conversational AI Tools

While Google was using AI to update what they showed to searchers, conversational AI tools like ChatGPT (launched in 2022) were gaining in popularity.

Over time, we’re learning that people aren’t using AI tools the same as traditional search engines. Queries on AI tools average 23 words compared to 4 on traditional search, and they frequently involve more follow-up queries than typical Google searches. And as we explored previously, Google use isn’t declining, so AI tools aren’t yet taking away users.

We’re also learning that people are starting to trust conversational AI tools more than brands. Even though AI tools can “hallucinate” and generate incorrect information in their answers, some users still trust them more than brands, due in part to poor experiences with self-serving branded content.

For marketers, it’s important to note that conversational AI tools are not yet a major source of website traffic. According to Ahrefs, these tools are the source for just 0.25% of a site’s total traffic on average, although those numbers did grow by 9.7x over a year. Our numbers for the CSU website are similar to those figures.

When AI tools cite sources in their answers, that’s what leads to those small amounts of tracked website traffic. And some content types are much more likely to be cited by AI tools compared to search engines: core website pages like product or about pages, documents like PDFs, and videos.

The Bottom Line for Conversational AI Tools

AI tools are not major drivers of website traffic (yet), but they’re quickly becoming one of the many ways that people learn about brands before visiting a site.

Therefore, optimizing for AI search is not only about optimizing for traffic. It’s about optimizing for visibility.

How Conversational AI Works

If we’re now trying to optimize for AI search, we should understand how these conversational AI tools work. Here’s a breakdown.

  • Conversational AI tools are powered by large language models (LLMs). These LLMs are trained on massive datasets that are scraped from content found across the web (or through partnerships with websites like Reddit or news publishers).
  • When generating answers, LLMs find patterns in their own datasets, then they use those patterns to calculate which words or entities are most likely to come after whatever preceded them. Some call this “spicy autocomplete.”
  • When an LLM can’t pull from its dataset to answer a question, it conducts an active web search to do so. Traditional SEO comes into play in these situations.

With all of this in mind, the goal of AI search optimization is getting your brand mentioned alongside the words you want to be mentioned next to, on websites that are likely to be used as sources by LLMs. Doing so improves your chances of being more visible in the dataset the next time an LLM scrapes content from across the web.

AI Search Optimization in 2026 and Beyond

Here’s where we get into what we should actually be prioritizing in our strategies moving forward.

For owned channels like your own website, traditional SEO tactics are still valuable, but the following are even more important in the era of AI search:

  • Simple, direct language and clear structures with headings, lists and internal navigation.
  • Standalone sections that can be extracted and sourced easily (FAQs, comparisons, pros and cons lists)
  • Content that goes beyond surface-level overviews.
  • Content that answers logical follow-up questions.
  • Unique insights from subject matter experts with deep knowledge that can’t be easily reproduced.
  • Personal perspectives and first-person language.
  • Original data and research.
  • Content with thorough citations that bolster credibility.

Off-site mentions can also significantly improve AI search visibility:

  • Content published via third party channels (YouTube, social media, etc.).
  • Earned media and sponsored content.
  • Websites and articles that publish rankings and reviews.
  • General brand mentions across the web (linked or unlinked), especially on sites frequently cited by AI tools (i.e., Reddit).

Strategic Considerations

What does all of that mean for us at CSU? The short answer is: we’re working on it. If you pay attention to our content, you’ll likely start seeing some changes, but we’ve got a lot of work to do to get there.

While we work on updating our strategies, I’m trying to keep these two considerations in mind:

  1. It’s still the early days of AI search optimization. Data and tracking tools for conversational AI tools are new and mostly unproven. And so far, many of the best practices for AI search overlap with the best practices for traditional SEO. Focusing on those fundamentals is still a smart way to spend your time.
  2. Due to the fact that (1) search engines are sending less traffic to websites, and (2) the degree to which AI tools will send traffic to websites in the future is uncertain, diversifying how your audience finds your content is even more important.

Sources