Search in 2026: Users Kept Both Google and AI — Now What?

The binary narrative — either Google or AI — turned out to be wrong. Research covered by Search Engine Journal shows that users are running both in parallel: querying ChatGPT for synthesis and exploration, then returning to Google for verification, local intent, and transactional tasks. Search volume on Google hasn’t collapsed. What’s collapsed is click-through to websites. That distinction matters enormously for how you architect a search visibility strategy right now.

The Dual-Engine Reality Changes Your Measurement Model

If users are using Google and AI simultaneously, you’re no longer competing for a single moment of attention — you’re competing for presence across two distinct decision surfaces. Traditional rank tracking already reflects the strain: BrightLocal’s Myles Anderson recently disclosed that Google’s anti-scraping updates have introduced significant latency into rank-reporting pipelines, forcing the company to rebuild core data infrastructure. This isn’t a minor technical inconvenience. It signals that Google is actively making its data harder to observe at scale — which, strategically, means your internal reporting cadence needs to adapt.

For Southeast Asian teams running campaigns across Thailand, Vietnam, and Indonesia, the implication is sharper: mobile-first users in these markets often skip desktop Google entirely and move between TikTok search, ChatGPT, and Shopee’s native search within a single discovery session. A rank-position dashboard alone tells you almost nothing about that journey.

Practical shift: supplement rank tracking with AI citation monitoring (tools like Profound or manual prompt auditing) and platform-native search analytics from Lazada, Shopee, and LINE Shopping. Map your content against all three surfaces, not just Google’s ten blue links.

AI Citation Is Winnable — But the Checkout Isn’t Ready

Here’s where the strategic gap is widest. Search Engine Journal’s Greg Jarboe makes the argument plainly: getting your product surfaced inside ChatGPT via agentic commerce protocols is, relatively speaking, the easy part. The hard part is what happens at checkout. When a user asks ChatGPT to buy something on their behalf, the handoff from AI recommendation to actual transaction is where most retailers are failing — broken product schema, mismatched inventory signals, payment flows that aren’t optimised for agentic handoffs, and return policy language that no LLM can confidently summarise for a cautious buyer.

Jarboe recommends three checks before connecting an agentic commerce protocol: confirm your structured data is complete and validated at the product level, audit your checkout flow for friction points that an AI agent — not a human — will encounter, and verify that your return and trust signals are machine-readable, not buried in PDFs or image-based FAQ pages.

For brands selling through Lazada or Shopee in Southeast Asia, the near-term version of this is ensuring your product listings are structured for both platform search algorithms and any AI-layer integrations those platforms introduce. Tokopedia and Shopee have both been testing AI-assisted discovery features — the brands with clean data architecture will get the early lift.

---

GEO Is Not a Separate Strategy — It’s the New Baseline

Generative Engine Optimisation used to be treated as an experimental layer on top of SEO. That framing is now outdated. If a meaningful portion of your audience’s discovery journey runs through an AI answer layer — and the evidence suggests it does — then being citable by those systems is table stakes, not a nice-to-have.

What makes content citable by LLMs differs from what makes it rank on Google, but the gap is smaller than the hype suggests. LLMs favour content that is: specific rather than general, structured with clear claims and supporting evidence, attributed to identifiable expertise, and consistent across multiple web sources. That last point is underappreciated — AI systems build confidence in a claim by finding it corroborated across independent pages. A single well-optimised pillar page is less citable than a coherent cluster of content making the same point from different angles.

For multilingual Southeast Asian brands, this creates a structural challenge: most content clusters are built in English, leaving local-language content thin and therefore uncitable in regional AI queries. Brands that invest in structured, authoritative Bahasa Indonesia, Thai, or Vietnamese content clusters now will have a meaningful head start when regional AI assistants — including LINE’s AI features and Grab’s in-app assistant — begin surfacing answers more aggressively.

What Dual-Engine Search Means for Local SEO

Local search intent hasn’t migrated to AI in any significant way yet. “Best mookata near Asok” still resolves on Google Maps, not ChatGPT. But the research showing declining click-through rates means that even local searchers are getting answers within the SERP — hours, reviews, menu items — without visiting a website. Google Business Profile completeness is no longer optional hygiene; it’s the primary content asset for any brand with a physical presence.

The more nuanced local play is ensuring that the information AI systems learn about your locations is accurate and consistent. When a user asks an AI assistant to recommend a pharmacy in Kuala Lumpur open on a Sunday, that system is drawing on web-crawled data — your website, your GBP, third-party directories. If those sources conflict on hours or location, the AI either hedges or omits you. Audit your location data across all sources with the same rigour you’d apply to product schema.

Key Takeaways

• Users are running Google and AI in parallel — your visibility strategy must be built for both surfaces simultaneously, not sequenced.
• Agentic commerce readiness is now a technical SEO problem: structured product data, machine-readable trust signals, and frictionless checkout flows are prerequisites, not differentiators.
• Multilingual content clusters in Southeast Asian languages represent an underexploited GEO opportunity — AI systems reward corroborated, structured authority across multiple sources.

The deeper question the dual-engine reality raises: if users are satisfied by AI-generated answers for synthesis tasks and Google answers for transactional tasks, where exactly does a brand’s owned website sit in that ecosystem? The brands figuring that out now — building content that is simultaneously citable, rankable, and conversion-optimised — are building an asymmetric advantage. The ones waiting for the dust to settle may find the landscape has already been claimed.

---

At grzzly, we work with mid-to-large brands across Southeast Asia on exactly this challenge — building search visibility strategies that span traditional SEO, AI citation architecture, and agentic commerce readiness. If your current approach was designed for one search engine and one kind of user journey, it’s worth a conversation. Let’s talk