5 Steps to Detect AI Overviews Eating Your Site Traffic

To tell if AI Overviews are eating your traffic, you must combine data from Google Search Console and Google Analytics 4. GSC’s Performance report can filter for ‘Generative AI’ impressions, but it does not show direct click-through rate loss. In GA4, you can create a segment to isolate traffic from pages known to have AIOs, using regular expressions to track changes in user behavior and conversion rates. This two-tool approach provides a directional, though imperfect, diagnosis of the impact from AI Overviews (AIOs).

This guide provides a step-by-step framework for diagnosing traffic changes specifically attributable to AIOs and adapting your strategy. By the end, you will be able to analyze your own data to measure the impact and re-optimize your content for visibility in an AI-driven search landscape.

Prerequisites

• Administrator access to your website’s Google Search Console (GSC) account.
• Editor or Administrator access to the corresponding Google Analytics 4 (GA4) property.
• A spreadsheet application like Google Sheets or Microsoft Excel for data logging.
• Basic familiarity with regular expressions (regex) is helpful but not mandatory.

Step 1: Identify Keywords Triggering AI Overviews

Before you can measure impact, you must identify which of your target keywords are generating AIOs. This has become a moving target since Google’s rollout, but a manual check is the most reliable starting point.

Create a list of your top 20-50 informational keywords (based on impression volume in GSC). Perform a Google search for each keyword in an incognito browser window. Log which queries trigger an AIO in your spreadsheet. Categorize them:

• High Impact: AIO provides a complete answer, making a click to your site unnecessary.
• Medium Impact: AIO provides a partial answer but still encourages a click for more detail. Your site is cited.
• Low Impact: AIO is present but irrelevant, or your site is a primary, prominent link.

Pay special attention to informational and question-based queries (who, what, why, how), as these are most susceptible to being answered directly by an AIO. Transactional queries (e.g., “buy running shoes size 11”) are less affected for now, but this could change. This initial audit gives you a qualitative baseline for the quantitative analysis to follow. Using a tool with AI keyword suggestion capabilities can help expand this list to find related queries that may also trigger AIOs.

Step 2: Analyze Performance in Google Search Console

Google Search Console is your primary source for understanding how AIOs affect your site’s visibility in search results. The platform has a specific filter for this, but its limitations are significant. Although Google has recently introduced a Generative AI features performance report in Google Search Console, it does not give the level of detail needed to identify where AI Overviews are replacing a user’s need to click through to a webpage.

Navigate to the Performance report in GSC. Click + New > Search appearance... and select Generative AI. This will filter your data to show only queries where your site appeared within an AIO. Analyze two views:

1. With the Generative AI Filter ON: Look at the queries, impressions, and clicks. A high impression count with a very low click-through rate (CTR) is a strong indicator that the AIO is satisfying user intent without a click. This is the essence of zero-click searches in the AIO era.
2. Compare Time Periods: Set your date range to compare the period after the AIO rollout (post-May 2026) with the same period before it. Filter for a specific URL or query you identified in Step 1. A significant drop in clicks and CTR, while impressions remain stable or increase, points directly to AIO impact.

Keep in mind that GSC data is aggregated and anonymized. It won’t tell you that a specific user saw an AIO and didn’t click. It provides trends, not certainties. This is where GA4 becomes necessary.

Step 3: Isolate Potential AIO Traffic in Google Analytics 4

While GSC shows you what happens on Google’s results page, GA4 shows you what the users who do click from AIO-affected queries do on your site. The key is to build a custom segment to isolate this traffic. There is no automatic “AIO traffic” source in GA4, so you must create a workaround.

First, you need a list of your site’s pages that rank for keywords that trigger AIOs (from Step 1). Let’s say you have 15 such pages. You will use regex to group them.

Create a GA4 Segment for AIO-Affected Pages:

1. In GA4, go to Explore and start a new Free-form exploration.
2. In the Segments section, click the + icon to create a new user segment.
3. Under Include users when, add a new condition: Session start.
4. Add a parameter to this event: page_path.
5. Set the condition to matches regex.
6. Enter a regex string that matches the URL slugs of your affected pages. For example, if your affected pages are /blog/what-is-x, /blog/how-to-y, and /blog/why-is-z, your regex could be ^/blog/(what-is-x
   

   how-to-y

   why-is-z)$.

Now you can apply this segment to any GA4 report. Compare the behavior of this segment against all users. Look for:

• Lower engaged sessions per user: Users might be arriving with more specific intent, finding their answer quickly, and leaving.
• Lower conversion rates: If the AIO skimmed the value from your page, users who click through may be less qualified or motivated to convert.
• Changes in user demographics or location: AIOs might surface your content to new audiences.

This method is a proxy, not a perfect measurement. It assumes traffic to these pages is primarily from AIO-affected queries, which is a reasonable starting point for diagnosis.

The goal is no longer just to get a click, but to be the definitive source cited within the AI Overview itself.

Step 4: Differentiate AI Crawlers from Human Users

Not all non-converting traffic is from humans who saw an AIO. An increasing portion of web traffic comes from AI bots, crawlers, and scrapers. Some of these, like Google’s own Google-Extended bot, are used to train generative models. Distinguishing these automated agents from human users is crucial for clean data.

Check your server logs for user agents associated with AI crawlers. Common ones include:

• Google-Extended (feeds Google’s Vertex AI and Bard)
• ChatGPT-User (OpenAI’s web crawler)
• CCBot (Common Crawl’s bot)
• PerplexityBot (for Perplexity AI)

If you see a spike in traffic from these user agents that corresponds with your overall traffic drop, you are likely ‘feeding’ AI models. This traffic will almost never convert and is characterized by single-page visits with 100% bounce rates. You can filter this bot traffic out in GA4 to get a cleaner view of human user behavior. Some sites have begun blocking these crawlers via their robots.txt file, a move that has significant and debated consequences for discoverability by AI systems.

Step 5: Adapt Your Content Strategy for AIO Visibility

The long-term solution isn’t just about tracking losses; it’s about adapting to survive and thrive. AIOs favor content that is clear, factual, well-structured, and authoritative. A shift in mindset is needed from just winning the click to winning the citation within the AIO.

Strategies to Optimize for AIOs:

• Build Topical Authority: Consistently publishing in-depth content on a specific niche signals expertise. This is a core principle of Topical Authority, which helps both traditional SEO and makes your domain a more trusted source for AI models.
• Use a Citation-Ready Format: Structure articles with clear headings, lists, and concise definitions. Answer the user’s question directly in the first paragraph. This Citation-Ready Format makes it easy for AI to extract your content and present it as a definitive answer.
• Add Unique Data and Insights: AIOs are excellent at synthesizing existing information. They are poor at creating novel insights. Include original research, unique data points, or expert commentary that cannot be found elsewhere.
• Focus on Complex Topics: Pursue topics that require deep expertise and cannot be summarized in a simple AIO. Think detailed case studies, multi-step tutorials, and nuanced comparisons.
• Diversify Traffic Sources: Reduce your reliance on informational query traffic from Google. Invest in social media, email newsletters, community building, and direct traffic. The rise of AI chatbots and zero-click searches makes a multi-channel strategy essential.

Troubleshooting Common AIO Analysis Problems

Diagnosing AIO impact is filled with ambiguity. Here are common issues and how to address them.

• Problem: My clicks are down, but GSC shows no ‘Generative AI’ impressions.

The GSC filter is not perfect and may not capture all instances. Your traffic drop could also be due to a core algorithm update, increased competition, or simple seasonality. Don’t assume AIOs are the sole cause without corroborating evidence.

• Problem: My traffic is down, but my engagement metrics are up.

This could be a positive sign. The AIO might be filtering out low-intent users. The traffic you are getting is smaller but more qualified, leading to better engagement from users who have a deeper need that the AIO couldn’t satisfy. This is a win, even if the top-line traffic number is lower.

There are workarounds. None of them are perfect, but together they give you a directional picture that can be pretty useful for understanding your AI visibility.

• Problem: I blocked AI crawlers, and my traffic dropped even more.

This is a known risk. While blocking crawlers stops them from scraping your content for free, it may also make your site invisible to the very AI assistants you want to be cited by. Some evidence suggests that being a source for models can have indirect ranking benefits. This is a strategic decision with no easy answer. As the web adjusts, many sites are finding that AI is eating website traffic.

Frequently Asked Questions

How can I get my content featured in AI Overviews?

Focus on E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness). Write in clear, simple language. Use structured data (schema markup) for lists, FAQs, and how-to guides. Answer the question directly and provide unique, valuable information that isn’t just a rehash of other top-ranking pages.

Do AI Overviews affect all types of websites equally?

No. Informational sites, particularly those in the YMYL (Your Money or Your Life) space like health and finance, and content-heavy blogs are most affected. E-commerce and local service sites are currently less impacted by AIOs for transactional queries, but this is subject to change as the technology evolves.

Is there any tool that can perfectly track AIO traffic loss?

Currently, no single tool provides a perfect, direct measurement of traffic lost to AI Overviews. The methods described in this guide, combining GSC, GA4, and manual checks, offer the most accurate directional analysis available. Third-party tools are emerging but rely on the same underlying data sources with similar limitations.

Will Google ever provide more detailed data on AIOs?

Google has a history of being opaque with data that could discourage ad spending. While improvements to the GSC Generative AI report are possible, it’s unlikely they will provide a clear ‘clicks lost to AIO’ metric. Site owners will need to rely on their own analysis for the foreseeable future.

Ultimately, adapting to AI in search requires a strategic shift. By implementing a process of consistent publishing and content optimization, you can build the kind of authority that AI systems are designed to recognize. To see how you can systematize this process, discover what you can do with the Keyworth automation platform.

Related reading

• Automate Content Creation for Organic Traffic.