SEO Pulse: Rankings and AI Search

Several developments in search marketing have raised questions about the reliability of ranking reports, ChatGPT referral traffic and the way search engines present links in AI-generated results. Recent updates from Google, OpenAI and Microsoft also offer new insights into AI search visibility and the future of Google Discover.

Here are the key developments and what they could mean for SEO professionals and content teams.

1. Google Ranking Reports Show Unusual Search Results

Between 13 September and 9 October 2026, rank-tracking company Nozzle found that 8% of the Google search results pages it monitored were missing most of the websites that would normally appear for those keywords. These sites subsequently returned in later checks.

The analysis covered 17.8 million search results pages collected through automated searches. The findings represent a significant increase from the 0.17% recorded between 20 August and 9 September. The proportion peaked at around 25% between 27 and 29 September before fluctuating between approximately 5% and 11% daily from 1 October onwards.

Nozzle described the unusual results as “poisoned” and suggested they could be connected to Google’s response to AI scraping. However, Google has not provided a statement confirming this explanation.

The impact on rankings was also notable. On affected results pages, a customer’s website failed to appear in the top ten for 58.6% of relevant keywords, compared with just 2.7% under normal conditions.

What This Means for SEO

Unusual fluctuations in ranking reports should not automatically be treated as evidence of a genuine decline in search performance.

Before reporting a sudden drop, compare current results with the websites that typically ranked for the same keywords before mid-September. Google Search Console can provide additional context by showing whether impressions and clicks for the affected pages and queries have also fallen.

Using both sources can help distinguish between actual changes in organic traffic and irregularities in the data collected by rank trackers.

2. ChatGPT Referral Traffic Falls Due to Missing Tracking Tags

A sharp decline in ChatGPT-attributed sessions appeared in Google Analytics 4 during October. Analysts linked the drop to a tracking issue in which some links within ChatGPT responses stopped including the utm_source=chatgpt.com parameter following the GPT-6 rollout.

Graphite chief executive Ethan Smith reported an approximate 45% reduction in ChatGPT sessions across websites with meaningful AI assistant traffic in GA4. Without the tracking parameter, some visits could be incorrectly attributed to Direct traffic rather than ChatGPT.

Adobe’s Cedric Huesler subsequently reported that OpenAI had confirmed the issue to him on 14 October and began rolling out a fix the following day.

By around 9 am Pacific Time on 15 October, Huesler observed that tagged ChatGPT clicks had returned to approximately 95% to 100% of the equivalent hour from the previous week across thousands of monitored websites. However, not all industry observers reported seeing the same recovery immediately.

OpenAI’s release notes were updated on 15 October, although they did not mention the specific tracking problem or its resolution.

What This Means for SEO

The apparent drop in ChatGPT traffic may reflect a measurement issue rather than a genuine loss of visitors.

The affected period was identified as 8 to 15 October, so SEO teams should make a note of it when reviewing monthly reports or comparing referral performance.

UTM parameters can help identify visits from ChatGPT, although not every visit will include a referral source or tracking tag. Reviewing both ChatGPT-attributed sessions and changes in Direct traffic can provide a more complete picture.

3. Google Changes How Links Appear in AI Search Results

Google is adjusting how links are displayed within AI Mode and AI Overviews, with some inline links now appearing higher in responses.

The company is also planning to test a section at the top of selected AI responses that directs users to official websites for particular tasks.

Other developments include the wider rollout of article carousels highlighting users’ Preferred Sources across additional locations and languages. Google has also begun testing the display of creator and publisher names on video results in AI Mode, linking these names to their Search profiles.

These changes could influence how users discover publishers and access supporting information while interacting with AI-generated answers.

What This Means for SEO

More prominent links may give websites additional opportunities to attract attention within AI search experiences. However, their visual placement does not necessarily translate into a higher reported position in Search Console.

Google’s reporting system assigns links within an AI Overview the same position as the overall overview. Consequently, moving an individual link higher within the response does not automatically change its recorded ranking.

Publishers should also be aware of the Preferred Sources feature. Websites must meet Google’s eligibility requirements for inclusion in relevant generative search features before they can qualify for this visibility opportunity.

Monitoring Search Console alongside changes to AI search presentation can help publishers understand how these developments affect their visibility.

4. Bing Introduces New AI Search Performance Insights

Microsoft has introduced a beta feature called Topics Portfolio within the AI Performance report in Bing Webmaster Tools.

The feature groups the queries for which a website appears in AI-generated answers into broader subject areas. It provides approximate data on the number of citations associated with each topic across websites, the number of sites cited and the number of related queries.

It also identifies opportunity queries where competing websites appear in AI answers but the user’s website does not. Citation volumes for these opportunities are presented as ranges rather than exact figures.

Topics Portfolio is only available when Bing has collected sufficient data for a website.

What This Means for SEO

The new feature gives content teams another way to identify gaps in their coverage and explore subjects where competitors are gaining visibility in AI-generated answers.

However, not every suggested query will be relevant. Some may focus on competitors’ brand names or topics that do not align with a website’s audience.

Microsoft advises treating the figures as directional indicators rather than precise measurements. They can be useful when comparing subject areas and identifying potential opportunities, but they should not be relied upon as exact citation counts.

5. Google Job Listings Point to AI-Generated Discover Content

Google’s publicly available job advertisements suggest that AI-generated content is becoming part of its plans for Discover.

A Staff Machine Learning Engineer listing for the Search Discover Feed Recommendation team describes work involving recommendations for AI-generated content based on users’ interests. It also references automated processes for producing content across proactive experiences, including Discover and Notifications.

A separate senior design role describes an ambition to develop Discover into a more proactive and personalised experience, involving collaboration with teams responsible for YouTube, Shopping, Chrome and Android.

Neither listing specifies when these changes will take place or how much AI-generated material might eventually appear in Discover.

What This Means for SEO

These job descriptions offer an indication of Google’s potential direction, but they do not confirm the scale or timing of future changes to Discover.

Publishers should continue monitoring their Discover performance and review the available Search Console reporting for generative AI features. According to Google’s documentation, the relevant Discover report measures impressions for AI features linking to websites.

Keeping track of these figures can help publishers understand whether their content is appearing in these experiences as Google’s approach develops.

The Main SEO Lesson: Verify Your Data

This week’s developments highlight a common challenge in SEO reporting: the numbers available to marketers do not always tell the whole story.

Rank trackers can encounter unusual search results, ChatGPT visits may lose their tracking parameters, and AI search reporting tools use different methods to measure visibility. Bing’s new topic metrics are approximate, while Google’s Discover AI reporting provides limited measurement details.

For more reliable reporting, SEO professionals should compare multiple sources rather than relying on a single metric.

Ranking fluctuations: Compare rank-tracking data with Google Search Console before attributing sudden changes to SEO performance.
ChatGPT referrals: Review both ChatGPT-attributed sessions and Direct traffic, particularly when investigating changes during the affected October period.
AI search visibility: Understand how Google and Bing define and report citations, impressions and link positions.
Discover performance: Monitor the available Search Console data while keeping in mind its measurement limitations.

The wider lesson is that search visibility is becoming more complex as AI features expand. Reliable analysis depends not only on collecting data but also on understanding how each platform measures and presents it.