Indonesia Singapore ไทย Pilipinas Việt Nam Malaysia မြန်မာ ລາວ
← Back to Blog

AI Search Is Changing How Clicks Work, Not Killing Them

Being cited in AI Overviews builds brand authority even when CTR drops — optimise for presence, not just clicks.

Editorial illustration of a figure fishing for clicks inside a giant AI search results page
Illustrated by Mikael Venne

AI Overviews cut CTR by 61% but clicks held steady. Here's what that means for SEO, GEO, and your traffic strategy in Southeast Asia.

The number that’s been circulating in SEO circles this week sounds alarming at first: AI Overview CTR dropped 61%. Before anyone drafts a crisis memo, the fuller picture is considerably more interesting — and more strategic.

The 61% CTR Drop That Isn’t a Disaster

Seer Interactive’s analysis, reported by Search Engine Journal, found that while click-through rates for brand-cited AI Overview results fell sharply, overall clicks didn’t collapse. The explanation is mechanical but important: impressions scaled dramatically faster than clicks. More people saw AI-generated answers; fewer of them needed to click through to get what they wanted. That’s not search dying — that’s search fulfilling its core promise more efficiently.

For brands operating in Southeast Asia’s high-velocity mobile environment, this distinction matters enormously. A user on a Grab commute asking a quick product question and getting an answer from an AI Overview that cites your brand is still a brand touchpoint. The session may not appear in your GA4 acquisition report, but the attribution gap doesn’t mean the influence gap is real. Citation is the new impression — and impressions have always preceded conversion.

Search Is Now a Task Completion Engine

Google’s latest feature releases, also covered by Search Engine Journal, make the directional intent unmistakable. Search is no longer primarily an index navigation tool. It’s increasingly an interface that completes tasks — comparing products, drafting responses, summarising policies — without handing the user off to a website unless genuinely necessary.

This rewrites the brief for content strategy. The pages that earn citations in AI-generated responses share a structural fingerprint: they answer discrete questions cleanly, present information in scannable, semantically unambiguous formats, and sit within domains that carry established topical authority. A Shopee seller’s FAQ page that precisely addresses “COD return policy for electronics in Thailand” will outperform a beautifully designed brand story page for that specific query type every time.

The implication for regional teams is that content governance — deciding which pages answer which intents, with what structure — now directly determines GEO visibility, not just traditional ranking position.


Traffic Channel Mix Is Fragmenting Faster Than Dashboards Reflect

Semrush’s analysis of billions of visits across more than 50,000 sites maps the broader channel shift playing out beneath the AI Overview story. Organic search’s share of traffic is redistributing, not disappearing, while direct traffic (a known proxy for brand strength and AI-assisted discovery) is holding or growing for established brands. Social and referral channels are performing differently by vertical, with some categories seeing meaningful AI-referred traffic that current UTM conventions don’t cleanly capture.

For Southeast Asian brands, this fragmentation is compounded by platform ecosystems that don’t behave like Western search funnels. LINE’s in-app search in Thailand, TikTok’s product discovery layer, and Lazada’s internal search all operate as semi-closed generative environments increasingly influenced by LLM-ranked content. A brand’s discoverability strategy needs entity-level consistency — the same brand name, product attributes, and category associations structured identically across platforms — because generative systems synthesise across sources before surfacing an answer.

Teams still optimising per-channel in isolation are measuring the wrong thing. The question isn’t “how is our organic traffic?” It’s “how consistently does our brand appear when a generative system answers a question in our category?”

Building for Citation, Not Just Ranking

The practical pivot for 2026 is moving from rank-centric SEO to citation-centric GEO. These aren’t opposites — strong entities still rank — but the optimisation logic differs. A few implementation realities worth addressing directly:

Structured data is non-negotiable at scale. Schema markup for products, FAQs, organisations, and reviews gives LLMs clean extraction paths. Brands running multilingual sites across English, Thai, Bahasa, and Vietnamese should audit schema consistency across language variants — inconsistencies create entity confusion that suppresses citation frequency.

E-E-A-T signals compound over time. Seer Interactive’s data implies that being cited in AI Overviews correlates with established domain authority. Brands that have been building genuine expertise signals — original research, author credentials, third-party citations — are pulling ahead in generative visibility. This isn’t a shortcut game.

Reporting infrastructure needs rebuilding. As Search Engine Journal notes, the surfaces businesses rely on for measurement haven’t kept pace with how search now works. Zero-click citations, AI-referred sessions miscategorised as direct traffic, and task-completion queries that never generate a URL visit are all invisible in most current dashboards. Teams should be experimenting with branded search volume trends and share-of-voice metrics as leading indicators rather than waiting for click data that may never arrive.


Key Takeaways

  • A 61% CTR drop in AI Overviews reflects impression volume growth, not audience loss — citation presence is itself a measurable brand signal worth tracking separately from clicks.
  • Search is now a task completion layer, meaning content structured for discrete, answerable intents earns generative citations; brand storytelling pages serve different jobs.
  • Southeast Asian brands operating across multilingual platforms need entity-level consistency — identical brand and product attributes across all surfaces — to register coherently inside LLM-powered discovery systems.

The more provocative question sitting underneath all of this: if generative engines increasingly complete tasks without transferring users to brand-owned properties, what does “owned media” actually mean in five years — and are we building the right assets for it?


At grzzly, we work with growth teams across Southeast Asia who are navigating exactly this shift — from rank-tracking orthodoxy to entity authority and citation strategy in LLM-powered search environments. If your reporting is still built around CTR and position data alone, there’s a conversation worth having. Let’s talk

Sneaky Grizzly

Written by

Sneaky Grizzly

Tracking the quiet revolution inside LLM-powered search — where brand mentions, structured semantics, and entity authority rewrite the rules of discoverability before most marketers notice.

Enjoyed this?
Let's talk.

Start a conversation