By Eric Richmond, CiteMetrix
Google has built a reporting surface for AI Overviews. But that reporting still cannot answer one of the most familiar questions in SEO:
Where did my link appear inside the answer?
That matters because position was the core currency of search visibility for the last twenty years. Rank 1, rank 3, page 2-these numbers gave SEO teams a shared language for visibility.
That language does not transfer neatly to AI-generated answers.
On September 14, 2026, Search Engine Watch reported that Google Search Console cannot show the position of an individual link inside an AI Overview. The report attributed the confirmation to Google’s John Mueller, who said that “position for these is hard to do in a way that makes it useful.”
That is a narrow claim about position data, not a claim that Search Console is useless for understanding AI search. Google’s reporting can still provide useful information about visibility at the feature or block level. But it does not tell you where your citation appeared within the answer.
That gap requires a different measurement vocabulary.
The practical problem with block-level reporting
An AI Overview is not a traditional list of ten blue links. It is a composed answer that may include several sources, recommendations, summaries, and follow-up paths.
If your URL appears inside that answer, the strategic meaning depends on how it appeared.
You may have been:
- The primary source supporting the answer
- A supporting citation among several sources
- A brand mentioned briefly without a link
- Absent while competitors were included
Those are four different visibility outcomes. They suggest four different actions.
A primary source may indicate that your content is shaping the answer. A supporting citation may mean your page is useful but not central. A passing mention may show that AI systems recognize your brand without directing users to you. Absence may reveal a citation gap or a competitor advantage.
A single block-level position cannot distinguish between them.
This is the asymmetry: the AI Overview is becoming a discovery surface, but the most familiar search metric was designed for a different kind of result.
The replacement vocabulary for AI visibility
If “position” is incomplete, what should marketers measure instead?
At CiteMetrix, we use a broader vocabulary for what we call brand knowledge governance: the cross-functional discipline of monitoring how AI systems understand, describe, and recommend a brand.
Here are seven metrics that make an AI answer more useful to analyze.
1. Inclusion
Were you present in the answer at all?
Inclusion is the most basic question. Did an AI platform mention your brand, cite your domain, or include one of your pages when responding to a relevant prompt?
Inclusion does not tell you whether the appearance was valuable. It tells you whether your brand entered the answer.
2. Citation
Did the AI platform credit your URL as a source?
A brand can be mentioned without receiving a link. Citation tracking separates those outcomes by identifying whether an AI platform credited a specific page.
The source matters, too. A product page, research report, comparison page, or help article may each play a different role in how AI represents your business.
CiteMetrix’s Source Pages analysis shows exactly which pages, yours and your competitors’, AI platforms cite. That gives content teams something more useful than a visibility total: a list of pages that are actually entering the answer.
3. Recommendation position
Were you named as a suggested choice, and how prominently?
For recommendation queries, inclusion is not enough. A brand listed first in “best tools for…” has a different opportunity than a brand mentioned at the end of a long answer.
Recommendation position measures your place among the options AI presents. It is especially relevant for SaaS, e-commerce, professional services, travel, finance, and other categories where users ask AI to compare or choose.
4. Prominence
How early and how centrally did you appear?
Prominence is the closest practical substitute for the internal position data that Search Console does not provide.
It can include observable signals such as:
- Whether your brand appeared in the opening answer
- Whether your source was attached to a central claim
- Whether you appeared in a recommendation list
- Whether the mention was buried in supporting context
- Whether your citation appeared before or after competitor citations
Prominence does not claim to recreate a hidden Google ranking formula. It measures the structure of the answer that users can actually see.
CiteMetrix’s ModelScore™ Mention Score weights citation rate, sentiment, and prominence across nine AI engines. Since Mention Score represents 45% of ModelScore, prominence is part of a larger view of whether AI systems are giving your brand meaningful attention, not merely mentioning it.
5. Sentiment
How were you framed?
A citation can be positive, neutral, or negative. A brand may appear frequently while being described as expensive, outdated, unreliable, or unsuitable for a particular use case.
That is why inclusion and citation counts need context.
CiteMetrix’s Model Sentiment and Brand Perception analyses track the language AI platforms use to describe your brand. This helps marketing, PR, and product teams see whether AI is carrying forward the positioning they intend.
6. Accuracy
Was what the AI said about you correct?
Visibility without accuracy can create risk.
An AI platform may cite your pricing page but state an outdated price. It may describe a discontinued feature as current or attribute a claim to your company that you have never made.
CiteMetrix’s Accuracy & Hallucination Detection checks AI responses against verified Brand Facts. It identifies five error types:
- Incorrect fact
- Outdated information
- Fabricated claim
- Misleading context
- Critical omission
Each issue receives one of three severity levels: Minor, Moderate, or Critical. Detection is included from the Starter plan, while Hallucination Watch automatically rechecks flagged issues daily until they are resolved on Professional and higher plans.
7. Competitor share of voice
Who else appeared in the answer?
Your visibility is relative. If an answer includes five competitors and your brand is missing, that is a different situation from an answer where you are the only recommended option.
CiteMetrix’s Share of Voice analysis provides head-to-head comparisons against named competitors across every monitored platform. It also powers Citation Gaps, which surfaces queries where your brand should be cited but is not.
This is often where the most actionable insight appears. The question is not only “Did we show up?” It is also “Who took the space we wanted?”
Why third-party measurement fills the gap
If the platform does not report internal placement, the only way to understand it is to observe the answers directly.
That observation needs to be:
- Systematic rather than occasional
- Repeated rather than based on one prompt
- Platform-specific rather than blended into one total
- Stored over time rather than treated as a single snapshot
Manually asking ChatGPT or Google one question can be useful for exploration. It is not a measurement program.
AI answers can change because of prompt wording, location, language, personalization, model updates, source changes, and simple response variation. A stable process is necessary before you can identify a meaningful trend.
A practical approach to measuring AI answer prominence
1. Build a stable prompt set
Start with the questions your customers actually ask.
Include:
- Category questions
- Brand questions
- Comparison questions
- Recommendation questions
- Problem-solving questions
- Questions where competitors are already visible
Keep the wording stable enough to compare results over time. You can add new prompts, but do not replace your entire set every week.
2. Sample more than once
One answer is an observation, not a conclusion.
Run the same prompt more than once and record whether the result is consistent. If your brand appears in one response and disappears in the next, that variation is itself useful information.
3. Record presence and prominence by platform
Google AI Overviews is a distinct surface. It should not be blended casually with ChatGPT, Perplexity, Claude, Gemini, or other conversational platforms.
CiteMetrix provides dedicated Google AIO tracking alongside its broader nine-platform coverage. This lets you see where your brand is visible and where it is not.
4. Track the answer, not just the number
A useful record includes:
- The prompt
- The platform
- Whether your brand was included
- Whether your URL was cited
- Your prominence or recommendation position
- The competitors included
- The sentiment of the mention
- Any factual errors
- The cited source page
This is also where the Content Change Tracker becomes important. CiteMetrix stores page snapshots on each scan, so a movement in citation or prominence can be compared with a content change in the same window.
5. Review trends over time
Do not treat a single ModelScore or citation result as a verdict.
Look for patterns:
- Are you being included more often?
- Are your citations moving closer to the central answer?
- Are competitors appearing in queries where you have a Citation Gap?
- Is sentiment improving?
- Are accuracy issues being resolved?
- Did a content update change which page AI cites?
AI visibility is not a one-time SEO audit. It is an ongoing brand knowledge governance program involving marketing, legal, product, and content operations.
From “What position are we?” to “How does AI represent us?”
The limitation reported by Search Engine Watch is useful because it exposes a larger measurement problem.
Search position answers a specific question: where did a result appear on a results page?
AI visibility requires several questions:
- Were we included?
- Were we cited?
- Which page was cited?
- Were we central to the answer?
- Were we recommended?
- How were we described?
- Were the facts accurate?
- Which competitors appeared instead?
No single number can answer all of those questions. But a measurement system can connect them.
That is the purpose of CiteMetrix. Its nine interconnected capabilities bring together Google AIO tracking, Source Pages, Citation Gaps, Share of Voice, Model Sentiment, Brand Perception, accuracy monitoring, Content Change Tracker, and reporting analysis.
When an executive asks how a number was computed, Report Analysis provides more than 10 report templates, PDF and PowerPoint export, shareable links, and scheduled delivery. You can move from a headline score to the underlying answer evidence.
CiteMetrix tracks nine AI platforms on every plan, with business plans starting at $79 per month. There are no extra apps or AI visibility add-ons required to assemble the measurement layer.
The next step
Google may improve its AI reporting over time. But until Search Console can show where an individual link appeared inside an AI Overview, “average position” is not a complete visibility metric for that surface.
Start with a stable prompt set. Measure inclusion, citation, prominence, sentiment, accuracy, and competitor share of voice. Then watch how those signals change as your content and brand facts change.
See how AI represents your brand and get your ModelScore → citemetrix.com
For background on Google’s AI search reporting, see Google’s documentation on AI features in Search. You can also explore how CiteMetrix calculates ModelScore and review the full analysis suite.
