What Website Data Tells Us About the Future of Search

To understand how AI is influencing organic visibility, I analyzed six months of traffic data (January to June 2025) across 10 different websites. These included a mix of national retailers, Canadian SaaS platforms, post-secondary institutions, B2B tech firms, and non-profit organizations. 

While this dataset is modest in size, it still paints an interesting picture of how AI is impacting SEO.

Here’s what the data told us and what it means if you care about increasing your visibility, traffic and sales in 2025 and beyond.

Blog glossary

Here are a few key terms that appear throughout the report:

• GenAI (Generative AI): Artificial intelligence that creates new content, such as text summaries, responses, or images, based on prompts and learned data patterns.
  
• GEO (Generative Engine Optimization): The practice of optimizing content for generative search tools (like ChatGPT and Perplexity), much like SEO for traditional search engines.
  
• AIO (AI Optimization): A broader term encompassing all techniques to increase visibility and conversions across AI-powered platforms.
  
• AI Overviews: Google’s AI-generated search summaries that aim to answer user queries directly in the search results, often reducing the need to click through to a website.

AI traffic by industry

While the total volume of traffic from AI tools is still growing, the growth rate and conversion behaviour are already showing promise across several industries.

What this tells us:

• A national retail chain received 1,301 visits from AI sources, up 574% over the previous period. 
• A Canadian SaaS company received nearly 4,000 visits from ChatGPT and Perplexity. 
• A college saw 3,182 visits from AI tools, a 348% increase.

Is AI actually driving website traffic?

Yes, but unevenly.

Some industries are already seeing noticeable traffic from AI search tools. Even when raw traffic numbers are low, AI-driven visits are growing fast. What’s important is not the total volume yet, but the growth rate. According to Semrush, AI traffic could surpass Google by 2028, potentially as early as 2026 in certain verticals.

On the other hand, smaller local websites or those lacking optimized content are seeing only a trickle.

Which AI tools send the most traffic?

This data matches what I’ve seen in other studies. ChatGPT is the dominant driver of AI referral traffic, but this could certainly change.

🧠 Some of these sources appear in GA4 as clean referrers, while others are misattributed as (direct) or (not set). But if you look closely at GA4 or referral logs, you’ll find them.

Are Google AI Overviews causing a drop in SEO traffic?

Yes, for some queries.

The Semrush study shows a clear drop in click-through rates as more users rely on AI summaries rather than clicking through to websites. That said, the impact varies:

• A B2B tech firm saw impressions up 20% but clicks slightly down 3%
  
• A non-profit’s impressions and clicks were both up 23%
  
• A college saw impressions up 31%, but clicks only increased by 1%
  

So far, brand strength and robust content strategies, including structured content and FAQ-style copy, seem to protect against major losses.

Are AI visits converting?

Not all traffic is equal, but AI traffic might be better.

In several cases, we found that visits from ChatGPT and Perplexity had lower bounce rates and longer session durations than traditional organic sessions. In the retail example above, users arriving from ChatGPT often landed directly on product pages or local store listings, suggesting strong purchase intent.

According to Semrush, AI-generated traffic converts at four times the rate of traditional SEO traffic. The reason? Users often complete their research inside the AI tool, so by the time they click through, they’re closer to taking action.

Changes in AI search vs. traditional search

AI search tools are more than just new traffic sources. These tools are changing how users behave.

This shift explains why AI-driven traffic often has lower bounce rates, longer sessions, and higher conversion rates. Users arriving from AI tools are often further along the buying journey.

What content gets featured in AI tools?

Across all accounts, these content types were most likely to generate AI referrals:

• FAQs and natural-language Q&A formats
  
• How-to guides, comparison charts, and checklists
  
• Original research, pricing tools, and calculators
  
• Store finders and service locations
  
• Domain expertise pages (e.g. “Do I need a will in Ontario?” or “What is SOC 2 compliance?”)

Tools like Perplexity and ChatGPT reference high-authority pages that answer questions clearly. Content that mimics the searcher’s question directly (not just optimized for keywords) is showing up more often.

Should we now “optimize for ChatGPT”?

Yes, but the playbook is still evolving.

That said, it’s starting to look a lot like the SEO playbook. The biggest shift is that instead of optimizing for Google’s algorithm, you’re optimizing for large language models (LLMs) that rely on clarity, structure, and trust.

Here’s what we’re doing for clients:

• Adding AI-friendly FAQs to core service pages
  
• Submitting LLMs.txt files to help models crawl and cite us
  
• Targeting mid-funnel questions that AI tools summarize
  
• Using schema and semantic HTML to clarify page intent
  
• Monitoring Perplexity and ChatGPT citations for visibility
  

If traditional SEO is about relevance and authority, then GEO is about clarity and trustworthiness. Again, the two overlap, but GEO requires content that is much easier for LLMs to understand, and can be deemed trustworthy.

This article is even formatted following these best practices. My goal here is to answer specific questions with unique proprietary data.

How can I measure GEO performance?

While tracking GEO is still a work in progress, we’ve been using referral traffic reports from GA4 and data from Semrush and waikay to measure GEO performance. 

Google is not being very helpful when it comes to reporting on AI Overviews/AI Mode data. While AI Overviews data flows into Search Console, there is no way to measure the number of impressions and clicks AI Overviews generates in Search Console. Google is lumping all the impression and click data together. Even Google’s AI Mode counts towards the totals you see in Search Console, according to Search Engine Land.

Here’s what you can measure:

• GA4
  • Referrals from chatgpt.com, perplexity.ai, gemini.google.com, etc.
  • (not set) traffic with long session durations and high scroll
    
• Search Console
  • Impressions and clicks in Google Search Console (however, Google is not labelling impressions/clicks as AI, rather lumping it in with traditional search clicks/impression data)
    
• Semrush
  • “AI Overviews” tracking feature shows how many search terms in their database you show up for
    
• Manual
  • Manual prompts in ChatGPT, Perplexity, and Gemini to see if your brand is cited
    
• Waikay
  • Discover how AI perceives your brand
  • Competitor comparisons
  • Analyze knowledge gaps

Most companies aren’t tracking this yet, which makes now the best time to get ahead.

At seoplus+, we started tracking AI referral traffic for all clients in Q2 2025. So, all of our SEO clients have insight into AI traffic trends for their websites.

Final thoughts

SEO isn’t dying, it’s evolving. The fundamentals still matter. But the way people find and consume information is changing fast. If your brand is not showing up in AI Overviews or AI tools like ChatGPT and Perplexity, you may already be losing traffic and leads.

The good news: you can apply much of what you already know about SEO (and what you’re already doing) to GEO. 

The bad news: if you wait too long, someone else will become the authority the AI chooses to cite.

If you want help measuring your AI visibility or building a GEO strategy? Reach out to seoplus+.

---