Michael Epps Utley {kind=image/jpg}
As Google continues to integrate AI into its search results, recent studies are shedding light on a key finding: AI-driven search remains highly unpredictable.
In a recent article, Matt Southern of Search Engine Journal highlights findings from a new study showing that Google’s “AI Mode” often returns completely different results for the exact same search, depending on when and how it’s performed. On the surface, this inconsistency might seem chaotic, but it actually presents a double-edged sword for digital marketers and SEO professionals.
The Pros and Cons of AI Variability
One potential benefit of this roll-the-dice approach is increased exposure for a wider variety of sources. Sites that might not typically rank on the first page of traditional search may get a chance to appear more prominently in AI-driven answers. This doesn’t mean every site will get a fair shake, but it means that worthy sources have a better chance of being seen.
But there’s a major downside too: ranking volatility. With AI Mode delivering different results session to session, businesses and content creators may find it difficult to measure success, replicate results, or even track performance in a meaningful way. The old playbook for SEO (where rankings were relatively stable and mostly measurable) doesn’t apply the same way here.
Despite the chaos, site authority still matters. The study notes that while AI Mode shakes up who gets seen, authoritative domains still have an edge. So, the fundamentals of SEO (quality content, backlinks, credibility, plus strong technical signals like site speed) are far from obsolete.
LLM Search: Still a Black Box
In a related piece, Roger Montti, also at Search Engine Journal, dives into the current state of LLM (large language model) search visibility tools. These tools are meant to help SEOs understand and optimize for AI-generated results, but right now, they're more guesswork than guidance.
Currently, using these tools is akin to trying to untangle a massive, complex knot: difficult, imprecise, and unlikely to become completely accurate anytime soon. That’s because AI search results—powered by models like Gemini or OpenAI's GPT—aren’t just pulling indexed pages. They're synthesizing, summarizing, and sometimes paraphrasing across multiple sources, often without citation.
What Should Marketers Do?
In this new landscape, here are a few key takeaways for marketers and SEO professionals:
Expect fluctuation. Don’t panic if your site disappears from one AI summary—it might reappear in the next session.
Focus on authority. Build credibility with original, high-quality content and authoritative backlinks. These signals still matter.
Stay flexible. Traditional ranking tools may offer less clarity, so lean into metrics like engagement, brand visibility, and direct traffic.
Experiment and observe. Keep testing queries and reviewing how AI search results evolve, especially in your niche.
AI-driven search is still in its early stages, and while it's shaking things up, it’s also creating new opportunities. The key is to stay informed, stay adaptable, and recognize that we’re all still learning the rules of this new game.
