GEO, AI Agents, and the New Rules of Search Visibility

The click-through rate data from France should make every regional search lead a little uncomfortable. Ahrefs tracked 963 domains across 28 days before and 9 days after Google rolled out AI Overviews in France on July 22, 2026 — and the most AI-exposed domains lost 23.1% of their CTR. That’s not a rounding error. That’s a structural shift in how search traffic flows, and Southeast Asia won’t be exempt for long.

What’s striking isn’t the number itself — it’s the speed. France went from zero AI Overviews to a double-digit CTR hit in under two weeks. Markets across Southeast Asia that are still optimising for traditional blue-link rankings are essentially building sandcastles at low tide.

GEO Is No Longer Theoretical — There’s Ad Revenue Attached

Generative Engine Optimisation moved from conference-circuit buzzword to commercial reality this month. Search Engine Journal has confirmed a live session with OpenAI on September 22, focused specifically on how brands can drive product sales and lead generation directly through ChatGPT — through a combination of GEO strategies and in-chat advertising placements.

This matters because it reframes the value proposition entirely. GEO is no longer just about brand mention frequency inside LLM responses — it’s now a paid media channel with conversion intent baked in. For marketers managing performance budgets across Lazada, Shopee, and Meta simultaneously, ChatGPT is about to earn its own line item.

The practical implication: brands that have invested in structured data, authoritative entity signals, and consistent brand voice across third-party publications are starting from an advantaged position. Those that haven’t are buying reach in a channel they don’t yet understand. The smart move right now is to run a GEO audit before committing budget — map where your brand is mentioned, in what context, and with what accuracy inside major LLMs.

Your Website Has a New Audience: AI Agents That Don’t Click

Crystal Carter’s five-part framework published on Moz this week is one of the more operationally useful things to land in the search space recently. The core premise: AI agents are already crawling your site, making decisions about your brand, and synthesising your content — and most websites are architecturally invisible to them.

The framework centres on a few concrete readiness signals. First, llms.txt — a plain-text file in your root directory that tells AI agents what your site is, what it covers, and what they should prioritise. Think of it as robots.txt, but for the LLM layer. Second, MCP (Model Context Protocol) integration, which allows AI agents to interact with your site’s functions, not just read its content. For e-commerce brands on platforms like Shopee or brands running their own D2C stores, this is the difference between an AI agent that can tell a user your price and one that can complete a transaction.

For Southeast Asian markets, the mobile-first architecture question is non-trivial here. Many regional sites have separate mobile URLs or app-exclusive inventory — neither of which AI agents navigate cleanly. Auditing your crawlability from an agent perspective, not just a Googlebot perspective, is now a legitimate technical SEO priority.

---

Local SEO Hasn’t Died — It’s Just Speaking a Different Language

The “near me” search intent isn’t fading. What’s changed is where and how that intent gets resolved. According to SEO.com’s recent analysis, local search queries are increasingly being processed inside AI-powered interfaces — Google’s AI Mode, ChatGPT, Perplexity — rather than traditional SERPs with map packs and star ratings front and centre.

This creates a specific problem for local and regional businesses: their Google Business Profile optimisation, which was the dominant local SEO lever for years, now needs to be complemented by signals that LLMs can actually parse. That means structured schema markup for local business data, consistent NAP (name, address, phone) information across every indexed mention online, and — critically — review content that includes specific, contextual language rather than generic five-star noise.

For multi-location brands operating across Thailand, Vietnam, Indonesia, and the Philippines, this compounds fast. Each market has its own platform ecosystem (LINE in Thailand, Zalo in Vietnam, Grab everywhere), and each of those platforms generates brand signals that mayor may not feed into LLM training data. The brands that win local AI search over the next 18 months will be the ones that treat every digital touchpoint — not just Google — as a source of structured, trustworthy entity data.

Where This Leaves Your Search Strategy Right Now

The honest read on this week’s signals: we’re in a 12-to-18 month window where the cost of inaction in GEO and AI agent readiness is still recoverable, but the gap is widening quickly. The France CTR data shows what happens when AI Overviews hit critical mass. The ChatGPT ad announcement shows where commercial intent is migrating. And the agent-readiness conversation shows that the technical surface area of “search” is expanding well beyond what most SEO playbooks currently cover.

For Southeast Asian brands, the question isn’t whether to engage with this shift — it’s whether you build the internal capability to navigate it or find partners who already live inside it.

Key takeaways:

• Run a GEO audit before allocating any budget to ChatGPT’s emerging ad placements — understand your current LLM brand footprint first.
• Implement llms.txt and audit your site’s agent-crawlability as a technical priority equivalent to Core Web Vitals.
• Extend local SEO signals beyond Google Business Profile to every platform that generates indexed, structured brand mentions in your target markets.

The search landscape is splitting into two tracks: one optimised for humans clicking links, one optimised for machines synthesising answers. The brands that assume those tracks converge neatly are going to be surprised by their Q1 2027 traffic reports. The real question is whether your search strategy is being built for the world that exists now, or the one that existed three years ago.

---

At grzzly, we work with growth and marketing teams across Southeast Asia on exactly this intersection — mapping LLM brand visibility, building GEO content frameworks, and auditing technical readiness for AI-powered search. If your search strategy hasn’t been stress-tested against the AI layer yet, that’s worth a conversation. Let’s talk