How AI Chat Is Changing The SEO Game

Search is changing in a big way. More and more, people aren’t scrolling through pages of links—they’re asking tools like ChatGPT or Claude a question and getting a single, stitched-together answer. That shift raises a simple but important question: if your website isn’t the thing people are clicking anymore, what does it mean to be visible online?

We asked our good friend and go-to SEO expert Josiah Casey how that new reality is playing out. We talk about what it actually means to show up in an AI-generated answer, why some brands get mentioned while others don’t, and what you can realistically do to improve your chances. Along the way, we get into the kinds of content that seem to resonate, how authority works when algorithms are doing the summarizing, and how to tell if any of it is paying off.

Is it as valuable to show up as a source in an AI-generated answer as it is to show up on a traditional search engine results page (SERP)?

Showing up in an AI-generated answer is at least as valuable as a standard organic search result today, and in some cases more so. Modern search results pages are so cluttered with ads, shopping features, and AI overviews that traditional organic listings are already getting pushed down. If someone is using a chat platform like ChatGPT or Claude, your Google rankings are irrelevant to them entirely.

That said, not all AI mentions are equal. Being listed alongside competitors as one option in a category is decent visibility, but being the recommended answer is a different level entirely. That kind of citation carries real weight, and the conversion data backs it up: traffic coming from AI tends to close at a higher rate, likely because those visitors have already done most of their research before they ever click through.

What practical steps can businesses take today to increase their chances of being cited by AI tools?

Start with an entity audit: make sure AI systems actually know who you are. That means a complete Google Business Profile and consistent name, address, and phone number across major directories. From there, identify the specific prompts you'd want to be cited for and go test them.

Most AI tools that pull live data are searching Google and scraping the top results, so if you show up in those organic results, you're more likely to appear in the AI's answer. That makes traditional SEO the most practical lever you have, just pointed at the right queries.

Are there specific types of content that perform better in AI-driven search?

Specific, detailed content tends to outperform broad topic coverage. FAQs, comparison guides, and best-of lists do particularly well, likely because they map directly to the kinds of questions people ask AI tools. Strong opinions from a credible source also seem to get picked up more than hedged, try-to-please-everyone content. The pattern is that AI tools favor content that resolves a specific question over content that simply addresses a topic.

How important are brand authority and backlinks for appearing in AI-generated answers?

Brand authority and backlinks both matter, but not necessarily in the way you'd expect.

Large Language Models (LLMs) appear to determine brand authority largely through citation frequency: how often your brand is mentioned across the sources they were trained on and actively retrieve from. Backlinks don't seem to be a direct input, but they influence your Google rankings, and since many AI tools are actively searching Google and pulling from those results, ranking well still feeds into whether you get cited.

There's also a source credibility layer worth paying attention to: AI tools tend to favor certain publications and websites when answering questions in a given category. Identifying which sources AI consistently cites in your industry, and working to get mentioned there, is probably more useful than chasing traditional link metrics.

How can we tell whether our content is being used by AI systems like ChatGPT or Google Gemini?

The most direct signal is your analytics: if an AI tool cited you and the user clicked through, you'll sometimes see referral traffic from platforms like ChatGPT or Gemini. It's incomplete data because attribution is increasingly unreliable, and not every AI link passes referral information, but it's the closest thing to ground truth you have.

Beyond that, most people turn to prompt tracking, which has its own limitations. We don't have access to what people are actually prompting these systems with, so identifying the right prompts to monitor is largely guesswork. The workaround is to run high-volume sampling on your best guesses and look for frequency patterns over time rather than trying to read any single result.