Google AI Overview and AI Mode: How to Optimize Your Site for AI Search

Open Google on almost any day and there’s a good chance you’ll see an “AI Overview”, or the “AI Mode” option, ahead of the organic results for many of your searches. It’s one of the biggest changes to search since Google launched, and it’s permanently changing how people find information and how SEO practitioners work.

Here’s the sobering part. SparkToro’s 2024 study found that more than half of Google searches in the US ended without a click to any website (zero-click search), and AI Overview, which answers questions directly on the results page, pushes that trend further. Many sites are losing traffic compared with the pre-AI era. At the same time, sites that adapt can gain visibility and brand awareness.

The problem is that a lot of SEO is still done the “ten blue links” way, focused on keyword density, exact-match anchor text and backlink counts, without adapting to a world where Google uses large language models to select and cite content. That’s one reason some sites that used to rank well are watching traffic slide.

This guide explains: (1) what AI Overview and AI Mode are and how they differ, (2) the real impact on organic traffic, (3) 12 techniques that help get your site cited in AI search, (4) tools for tracking AI citations, and (5) five SEO practices that no longer work in the AI era.

AI Overview vs AI Mode: what’s the difference?

First, you need to understand the difference between these two Google features, because they affect your strategy differently.

Google AI Overview (formerly SGE)

AI Overview shows an AI-generated summary at the top of the results page when Google decides a query suits an AI summary. It launched officially in May 2024, having been tested in beta as the Search Generative Experience (SGE).

Characteristics of AI Overview:

Appears as a panel at the top of the SERP
Summarizes the answer in a few short paragraphs
Includes citations linking to a handful of sources used to build the answer
Has a “Show more” option for extra detail
Only appears for queries Google considers suitable, mostly informational, how-to and comparison queries

AI Overview now appears on a large share of informational and educational queries, and that share keeps growing.

Google AI Mode (launched 2025)

AI Mode is a newer feature Google rolled out in 2025: a whole new search experience that users reach through the “AI Mode” tab on the results page or a dedicated URL.

Characteristics of AI Mode:

A conversational interface, similar to ChatGPT
Handles deeper, multi-step reasoning
Lets users ask follow-up questions in the same thread
Powered by Google’s Gemini models
Includes citations, placed inline within the text
Supports multimodal input (images, video, voice)

The key difference:

AI Overview appears inside the traditional SERP (anyone doing a normal search may see it)
AI Mode is a separate experience (users have to choose to enter it)

The impact on organic traffic

Here is what the research and the broader industry are showing.

Zero-click search keeps rising

The trend that multiple sources agree on:

Before AI: zero-click already existed thanks to Featured Snippets and Knowledge Panels
SparkToro (2024): more than half of US searches produced no click
Today: AI Overview and AI Mode push that share higher still

In plain terms, many people who search on Google never click through to a website. They get their answer from AI Overview, a Featured Snippet, a Knowledge Panel or AI Mode.

Click-through rate for position 1 is falling

The pattern being observed:

Position 1 on a query without an AI Overview: CTR roughly as before
Position 1 on a query with an AI Overview: CTR noticeably lower
If your site is cited in the AI Overview: you tend to lose fewer clicks and gain brand visibility

The conclusion: a site cited in AI Overview loses less traffic than one that isn’t. So the goal is no longer just “rank number one”. It’s “rank well and get cited in AI Overview”.

Brand search is becoming more important

Another notable shift is that branded search (people searching your brand name directly) matters more, because users:

See a brand mentioned in AI Overview and then search for it
Research with ChatGPT or Claude first, then search Google for a brand that caught their eye
Encounter brands through AI answers in a way that builds awareness rather than immediate clicks

That makes brand-building one of the most important SEO strategies today, not just keyword optimization.

10 SERP features that feed AI search answers

AI Overview and AI Mode don’t create answers out of thin air. They draw on information associated with these SERP features:

Featured Snippets: a major source for AI Overview on many queries
People Also Ask: used to generate follow-up questions
Knowledge Panel: entity data
Local Pack: location-based queries
Reviews and ratings: product and service queries
Video carousel: relevant videos
Image pack: supporting images
News box: trending topics
Top Stories: time-sensitive queries
Forums and discussions: Reddit and Quora citations have grown sharply

If your site already wins these SERP features, your odds of being cited by AI Overview go up. See also our AI SEO guide.

How AI search chooses its sources

From industry observation and published studies, AI search (Google AI Overview, Perplexity, ChatGPT Search) tends to favor sources on the following factors.

Factor 1: Authority

AI search favors high-authority sites, as reflected in:

Backlink profile (quantity and quality)
Domain authority metrics
Brand mentions across the web, linked or not
Domain history
Topical authority (being an expert on the subject)

Factor 2: Content quality

E-E-A-T signals (Experience, Expertise, Authoritativeness, Trustworthiness)
Original research, data and statistics
In-depth coverage of the topic
Author credentials
Last-updated date

Factor 3: Direct answer format

Clearly formatted answers, such as “X is…” followed by a definition
Bulleted lists, numbered lists, tables
Answers that are “snippet-able” and can be lifted as-is
Question-and-answer structure

Factor 4: Freshness

Recently updated content, especially for time-sensitive topics
Accurate last-modified information
Dates in schema

Factor 5: Structured data

Schema.org markup (Article, FAQ, HowTo, Product)
Breadcrumb schema
Author markup (who wrote it, with credentials)

Factor 6: User engagement

How long people stay
Click-through rate from traditional results
Whether people return

12 techniques to get cited in AI Overview

These are the techniques we use to make sites more citable in AI Overview. None of them can guarantee a citation, but together they make your content as easy to choose as possible.

1. Direct answer format that AI can lift

Write content in a shape AI can easily “extract” an answer from:

<h2>What is X?</h2>
<p><strong>X</strong> is [a clear one- or two-sentence definition]. X works by [a short explanation].</p>

Avoid vague openings like “Many people wonder what X is… some think…” because AI can’t use them.

2. Structured data (FAQ, HowTo, Article)

Implement the following schema types fully:

Article schema: every blog post
FAQ schema: FAQ pages and articles with an FAQ section
HowTo schema: tutorials and step-by-step articles
Product schema: e-commerce
Organization schema: the homepage
Author markup: every article with a named author

Structured data helps Google understand your content when selecting sources. Our Schema Markup guide covers implementation.

3. Clear E-E-A-T signals

E-E-A-T = Experience + Expertise + Authoritativeness + Trustworthiness

Ways to show E-E-A-T:

Author bio with clear credentials (for example, years of SEO experience and relevant certifications)
About page explaining who you are in detail
Trust signals: privacy policy, terms, contact details, address
Citations: reference research and authoritative sources
Case studies with real, verifiable figures

4. Conversational tone

AI search tends to favor content written in a natural, human way over stiff “SEO copy”, because language models are trained largely on natural language.

Example:

Stiff: “Cloudflare CDN reduces TTFB by leveraging global edge POPs…”
Natural: “If you put your site behind Cloudflare’s CDN, your TTFB drops because content is served from edge locations close to your visitors…”

5. Citable statistics

Include clear, sourced figures that AI can cite. AI likes specific numbers with an attribution.

Example:

Weak: “Lots of websites use Cloudflare.”
Strong: “Roughly a fifth of all websites use Cloudflare, according to W3Techs.”

6. Topic depth and breadth (topical authority)

AI search values topical authority: sites that cover a subject both deeply and comprehensively.

How to build topical authority:

Write pillar content covering the main topic in depth
Write cluster content covering the subtopics
Link every cluster page to its pillar and vice versa
Cover the bulk of the questions users ask about the topic

7. Strategic internal linking

Internal links help Google and AI understand how topics on your site relate, which strengthens topical authority.

Best practices:

Link each article to several related articles
Use descriptive anchor text, not “click here”
Link from informational pages to commercial pages (services, products) where it helps the reader

8. Brand mentions (linked and unlinked)

AI search gives weight to brand mentions, both linked and unlinked.

Ways to earn brand mentions:

Digital PR (coverage on news sites)
Guest posting (a mention plus a link)
Podcast appearances
Speaking at events
An active social media presence

9. llms.txt and llms-full.txt

These are emerging files that help AI crawlers understand your site:

llms.txt: tells AI what your site is about and where the main pages are
llms-full.txt: the full content of your site in plain text for AI to use

Support varies by provider, but they’re cheap to create. An example llms.txt:

# Southern Whale

> SEO and web development freelance team in Thailand
> Specializing in Astro, Next.js, WordPress, multilingual SEO

## Services
- [SEO Services](/en/services/seo/)
- [Web Development](/en/services/web-development/)
- [SEO Audit](/en/services/seo-audit/)

## Blog
- [AI SEO Guide](/en/blog/ai-seo-thailand/)
- [Core Web Vitals Guide](/en/blog/core-web-vitals-guide/)

10. Schema markup on every page

Beyond Article, FAQ and HowTo, add:

BreadcrumbList: every page
WebSite: the homepage
Organization with sameAs (links to social profiles): the homepage
Person: author pages
Service: service pages
LocalBusiness: local businesses

11. Author bio and credentials

Every article should have:

A real photo of the author
Their real name
Title or position
Credentials (qualifications, certifications, years of experience)
A link to an author page
Author markup

12. Original research and data

AI search likes to cite sources with original information, because that makes them authoritative.

Ways to create original research:

Survey your customers and publish the results
Case studies with real figures
Analyze public data and share insights
Industry benchmark reports
Run experiments and share the results

5 SEO practices that are “dead” in the AI era

Practices that worked a few years ago and no longer do in AI search.

1. Exact-match anchor text

Backlinks with anchor text matching your keyword used to help rankings a lot. Now context matters more than the anchor itself, and too much exact-match anchor text looks like spam.

2. Keyword density

“This word must appear X times in the article” is dead. Search systems understand semantics and context far better than word counts.

3. Quantity over quality in backlinks

“Whoever has more backlinks wins” is over. A handful of links from genuinely authoritative sites is worth more than piles of links from a PBN.

4. Optimizing single pages without looking at the whole site

“Optimize this page for this keyword” is no longer enough. You have to think about the whole site in terms of topical authority and site architecture.

5. Spun content (rewriting other sites’ content)

Completely dead. Google and AI systems are good at spotting spun content, and it’s exactly the kind of low-value material Google’s helpful content systems are designed to demote.

How to check whether your pages are cited

Rather than relying on anyone’s claims, check your own pages directly:

Pick the queries your key articles target, such as “what is [your topic]”.
Search them in Google, Perplexity and ChatGPT Search and note whether your page appears as a cited source.
Repeat monthly and keep a simple log, because results shift over time.

When pages do get cited, they typically share the traits covered above:

A direct answer format (for example, “Cloudflare is…”)
Citable, sourced statistics
Thorough coverage of the topic
Author credentials and E-E-A-T signals
Solid internal linking and topical authority
Complete schema markup (Article, FAQ, author)
llms.txt in place

Tools for tracking AI citations

A new set of tools has emerged for tracking whether your site is cited in AI search.

1. Brand Radar (Ahrefs)

Tracks brand mentions in AI responses
Covers ChatGPT, Perplexity, Gemini, Google AI Overview and more

2. Profound

AI search analytics platform
Tracks citations and share of voice in AI responses
Aimed at larger companies

3. Otterly.ai

AI search monitoring tool
Often used by smaller businesses

4. Peec AI

Tracks citations across AI Overview, ChatGPT and Perplexity
Alerts you to new citations

5. AthenaHQ

AI SEO analytics
Focused on conversational AI optimization

Check each vendor’s current pricing and coverage before committing, as this category changes quickly.

5 common AI SEO mistakes

1. Treating AI SEO as a “new trick”. It’s really a mindset shift from keyword-focused to question-focused content.

2. Neglecting traditional SEO. AI SEO doesn’t replace traditional SEO. You need both.

3. Prioritizing quantity over quality. A few genuinely useful articles beat dozens of average ones.

4. Not tracking AI citations. If you don’t know whether you’re being cited, you can’t judge the return.

5. Treating AI SEO as one-and-done. You need to keep updating content and monitoring citations.

Frequently asked questions

1. Does AI Overview appear in every country? Google has rolled AI Overview out to many countries and languages, including Thailand, and coverage keeps expanding. Whether one appears still depends on the query.

2. Do I need to change my whole SEO strategy? No. Traditional SEO still matters (technical SEO, content quality, backlinks), but you need to add AI optimization on top: schema, llms.txt and direct-answer formatting.

3. Will AI search replace traditional Google search? Nobody knows for sure. AI experiences are likely to keep growing for informational queries, while transactional and commercial searches still lean heavily on the traditional results page for now.

4. ChatGPT Search vs Google AI Overview: what’s the difference?

ChatGPT Search is conversational and suited to deeper questions
Google AI Overview appears on the normal results page and is suited to quick answers

5. Why does Reddit show up so often in AI Overview? Google signed a deal with Reddit in 2024 giving it access to Reddit content for its AI products, and Reddit threads have become a prominent source in results.

6. Should I post on Reddit for AI SEO? Genuinely helpful answers in subreddits relevant to your industry can raise your brand’s visibility, linking to your site only where it truly helps. Self-promotion is quickly removed, so it has to be real participation.

7. Is llms.txt necessary? It’s worth doing because it’s cheap and low-risk, but it isn’t yet an official standard and support varies between AI providers. Treat it as a helpful extra, not a substitute for good content and crawlability.

8. Will using AI to write content hurt rankings? It depends on quality. Google doesn’t ban AI content outright; helpfulness is the test. AI-written content that isn’t edited and fact-checked tends to fail because the quality is low.

Summary: AI SEO is the future, but traditional SEO is still the foundation

AI Overview and AI Mode are permanently changing how Google and AI search work: more zero-click searches, lower CTR for top positions when an AI Overview appears, and branded search becoming more important. This isn’t a trend. It’s a shift in era.

The way to win is to do traditional SEO well as your foundation, then add the 12 AI optimization techniques in this guide: direct answer format, structured data, E-E-A-T, conversational tone, citable statistics, topical depth, internal linking, brand mentions, llms.txt, schema markup, author bios and original research.

At Southern Whale, our freelance team works on both traditional and AI SEO. We can’t promise any page will be cited, because nobody controls that, but we’ll do our best to make your content as citable as possible. If you want your site to compete in AI search, contact us about an AI SEO audit, or see our SEO service.

Starting early matters: sites that ignore AI search risk steadily losing visibility, and the longer you wait, the harder it is to catch up.