Short answer
AI Overviews summarise a single query above the normal results. AI Mode replaces the results with a conversation and breaks your question into many related searches. Search Console reports both as one number, so you cannot tell them apart in your data. They tend to reward different things.
Most advice treats these as one surface. That is understandable, because Google’s own report merges them.
But they work differently, and optimising for one does not automatically win the other.
The two side by side
| AI Overviews | AI Mode | |
|---|---|---|
| Where it appears | Above normal results | Its own tab, replacing results |
| Organic links shown | Yes, alongside | No traditional results |
| When it triggers | Google decides per query | Whenever the user chooses it |
| Conversation | Single answer | Follow-up questions |
| How it searches | Retrieval for the query | Query fan-out |
| Personalisation | Limited | Uses account signals |
Both run on Gemini and both are grounded in Google’s search index. Google has said its AI search experiences still rely on its core ranking systems, so ranking well remains the precondition for both.
What query fan-out actually means
This is the mechanism that separates them.
Google describes AI Mode as breaking a complex question into multiple related searches across subtopics and sources, then combining the results into one answer.
So a question like “why did my traffic drop after an update” might be split into searches about update dates, technical causes, seasonality and recovery time. Each of those sub-searches has its own best answer.
Why they tend to cite different pages
Studies consistently find low overlap between what the two cite for the same question.
Treat any precise overlap percentage with caution. Published figures vary widely between studies and depend heavily on how each one sampled queries. The direction is consistent even where the numbers are not.
The reason follows from fan-out. An AI Overview answers one query, so it leans toward the page that best answers that query. AI Mode answers several sub-queries, so it can draw on a different page for each.
What each one tends to reward
This is informed reasoning from how they work, not a published formula. Neither surface discloses its selection criteria.
- AI Overviews favour a single liftable answer. A clear, direct response in the first sentence under a question heading.
- AI Mode favours topical depth. Covering a subject and its related sub-questions across several connected pages, so you can be the best answer to more than one fan-out search.
The overlap is large, though. Accurate, current, well-structured content that answers clearly helps with both. You are not choosing between them.
Why topic clusters matter more now
A single strong page can win an AI Overview. It is less likely to cover every angle a fan-out search explores.
This site is built as one example. The traffic drop diagnosis series is thirty connected pages, each answering one specific situation, all linked from a central triage page. That structure gives a fan-out search a relevant page for many of its sub-questions.
The principle is cluster coverage, not page count. Thirty thin pages would hurt you. Thirty genuinely distinct, useful ones help.
The reporting problem
Search Console’s Generative AI report combines AI Overviews, AI Mode and Discover AI into a single impressions figure.
- You cannot separate them in the report.
- You cannot see clicks, queries or position.
- A rise could be either surface, and the strategy differs.
The only practical check is manual. Search your key questions in both, note which of your pages each cites, and repeat monthly. Reading the report itself is covered in generative AI impressions dropped in Search Console.
For platforms outside Google entirely, see how to get cited by ChatGPT, Perplexity and Claude.
Checking both surfaces for your key questions
- List ten questions your best pages answer, written the way a customer would ask them.
- Search each one in normal Google and note whether an AI Overview appears and which sources it cites.
- Ask the same question in AI Mode, then one natural follow-up, and note the sources again.
- Record it in a simple sheet: question, date, cited in Overview, cited in AI Mode, which page.
- Repeat monthly with the same questions, so the comparison means something.
Use a logged-out browser where you can. AI Mode uses account signals, so your own search history can shape what you see.
Patterns matter more than single results. Cited in Overviews but never in AI Mode usually points at depth: one strong page, but nothing covering the related questions a fan-out search explores.
Repeat the check after any significant content change. When a new topic cluster goes live, the AI Mode column is where the effect should show first, because fan-out searches reach the related pages.
Reporting AI Mode traffic
Search Console has counted AI Mode activity within its ordinary web search totals since mid-2025. That means a change in AI Mode visibility can hide inside normal-looking numbers.
Use the monthly question check above to see the change directly, and the generative AI reporting guide for what Search Console now separates out. Report the two surfaces separately where you can, because they reward different kinds of page.
Keep the two reports side by side over several months. A gap that widens or narrows over time says far more about your content than any single month, and it shows whether changes you make are reaching the surface where you were weakest.
Cited in one surface but not the other?
That usually means a single strong page without the surrounding depth, or depth without a clear direct answer. Send your Search Console access and within 48 hours I will tell you which gap you have.
Get free recovery audit