A Reddit question — “Is SEO still enough, or do we need to think about GEO?” — got a short, clear answer from Google Search Advocate John Mueller: labels don’t matter as much as reality. If your business depends on referred traffic, Mueller wrote, “consider the full picture, and prioritize accordingly.”
That blunt, practical tone is worth repeating: AI is not a fad, but whether you should reorganize teams or rewrite your strategy depends on how your users actually find you.
What the Googlers said — and why it’s different from buzz
Mueller didn’t endorse a new discipline called GEO (generative engine optimization) or insist that SEO is obsolete. Instead he urged businesses to look at usage: “Be realistic and look at actual usage metrics and understand your audience (what % is using ‘AI’? what % is using Facebook?).”
Danny Sullivan, meanwhile, has been making the same point in public forums and on the Search Off the Record podcast: hiring a GEO or buying an “AI-optimization” tool is similar to hiring an SEO or buying an SEO tool. Know the fundamentals first. If you don’t understand how Google sees your site, you can’t effectively evaluate a consultant or a product.
Those two threads — practical measurement and fundamental literacy — are the throughline. The labels (GEO, AEO, SEO) are shorthand, not strategy.
The traffic reality: AI referrals exist, but they’re small for most sites
Industry checks back up Mueller’s caution. Current estimates show AI assistants and chat-based engines sending measurable but modest referral shares for the average publisher (ChatGPT referrals land in the tenths of a percent range for many sites). That’s growing, but it’s not replacing search overnight.
So the right question isn’t “Is GEO real?” but “How much of my audience is actually using AI assistants today, and how fast is that changing?” That answer will tell you whether to experiment lightly, dedicate a team, or leave most resources on traditional SEO and product channels.
Practical steps for teams that want to be ready (without chasing every buzzword)
- Audit your analytics. If AI referrals show up, track which pages are being cited and what outcomes (clicks, purchases, signups) follow. Treat AI like any other channel and measure ROI.
- Prioritize by impact. If search still drives the lion’s share of revenue, keep fundamentals strong: crawlability, helpful content, technical health. If AI is already meaningful for you, run focused experiments to surface what types of content are most likely to be cited.
- Ask the right questions when hiring or buying. Do the consultants/tools follow Google’s public guidance? Can they explain how improvements map to real user outcomes? Beware of vendors who obsess over third-party domain grades or “spam scores” as the only success metric.
- Keep writing for people. Both Mueller and Sullivan emphasize this: tools that promise “write-for-LLMs” shortcuts can lead you away from what actually converts humans.
Tools, agencies and the danger of metric fixation
Tools have gotten smarter, but they can also mislead. Some platforms highlight domain scores or micro-ranking signals that don’t line up with how Google evaluates helpfulness. Sullivan points out a common complaint — high domain scores but low real-world performance — and reminds teams that those vendor metrics aren’t Google’s metrics.
If you buy a tool or hire a firm, make sure you understand how they measure success, what changes they will make, and how they avoid tactics that could trigger a manual spam action.
A small ecosystem that’s changing fast
AI features are proliferating across products and contexts: from Google’s agentic booking experiments to deeper integrations that let models search your Drive and Gmail — these shifts change discovery pathways and attribution. If your strategy touches search or assisted answers, it helps to watch where those platforms are headed: for example, Google’s agentic booking efforts show how AI can become transactional rather than just referential, and Gemini’s expanding workspace integrations point to new kinds of reach outside traditional search.Google’s AI Mode adds agentic booking and Gemini’s Deep Research linking into Gmail and Drive are examples of that trend. Even platform-level partnerships matter: Apple’s decision to use a custom Gemini model for Siri highlights how major ecosystems are adopting these models in ways that can reshape referrals.Apple to use a custom Google Gemini model to power Siri
Treat those developments as context, not as an automatic reason to rip up your playbook.
Mueller’s short, practical advice is easy to forget in the noise: measure first, prioritize where it moves the needle, and don’t confuse product names with strategy. For most teams that means keeping SEO fundamentals in place, running small experiments with AI channels where the data suggests opportunity, and hiring or buying only after you understand the guardrails and the likely payoff.