If you've spent any time looking at your Google Search Console analytics recently, you might be feeling a familiar sense of dread. Traffic patterns are shifting. Click-through rates on previously rock-solid informational queries are wobbling.
The culprit? Google's AI Overviews (formerly SGE), alongside the rapid rise of answer-engine platforms like Perplexity, ChatGPT, and Claude.
Over my nearly 30 years in digital marketing, I've seen the industry declare "SEO is dead" more times than I can count. Every major algorithm update—from Panda to Penguin, and now to Generative AI—sparks the same panic. But let me be entirely clear: SEO isn't dying; it is fundamentally evolving.
If you are worried that AI Overviews are killing your organic traffic, you are asking the right question. But the answer isn't to abandon search. The answer is to pivot your strategy towards Large Language Model Optimisation (LLMO).
Here is exactly what is happening to organic search and how you need to adapt your strategy today.
What Are AI Overviews Doing to Search Traffic?
AI Overviews are designed to give users a comprehensive answer directly at the top of the search engine results page (SERP), without them needing to click a blue link.
For top-of-the-funnel, purely informational queries—like "what is a backlink" or "how to calculate ROI"—zero-click searches are absolutely increasing. If a user's question can be answered effortlessly by an AI summary, they will not visit your website to read a 2,000-word article padded with fluff.
However, AI Overviews are struggling to replicate three crucial elements:
- First-hand experience and unique data.
- Deep, transactional trust.
- Complex, nuanced problem-solving.
Traditional SEO focused almost exclusively on ranking for keywords. Today, success requires optimising for entities and AI recommendations. You want ChatGPT, Claude, and Google's AI to recognise your brand as the definitive authority in your space so they cite you in their responses.
How to Pivot Your Strategy for the LLM Era
To thrive in a landscape dominated by generative AI engines, you must shift your mindset from "optimising for human-read SERPs" to "optimising for AI ingestion". Here is the blueprint.
1. Shift Your Focus to Information Gain
AI models are trained on the existing public internet. If your content merely regurgitates what has already been said, an LLM has no reason to cite you over Wikipedia or a massive publisher.
You need to produce content with high Information Gain. This means bringing net-new value to the internet.
- Proprietary Data: Share case studies, original research, and statistics that only your business possesses.
- Expert Opinions (Experience & Expertise): Lean heavily into Google’s E-E-A-T guidelines. An AI can summarise facts, but it doesn't have 30 years of battle-tested industry experience. Inject strong, distinct viewpoints into your writing.
2. Structure Content for AI Extraction
LLMs love structure. When scraping the web to formulate answers, AI models look for clear semantic signals to understand context.
- Use Clear Headings: Your H2s and H3s should be direct questions or crystal-clear statements.
- Format with Lists and Tables: Answer engines frequently extract bullet points and tabular data to present to users. If you are comparing two products, put the comparison in a clean, easily readable table.
- Provide Direct Answers: Adopt the "Inverted Pyramid" style of journalism. Answer the core premise of the article in the first two paragraphs concisely, and then elaborate on the details further down.
3. Prioritise Brand Mentions and Digital PR (AI-Enhanced Link Building)
How does ChatGPT know which software to recommend for enterprise accounting? It doesn't rely on a traditional PageRank algorithm; it relies on the consensus of its training data.
If your brand is constantly mentioned on highly authoritative sites, podcasts, and industry forums alongside specific keywords, the AI connects your brand (the entity) to the solution.
- Move away from low-quality, high-volume directory links.
- Focus on AI-driven Digital PR and securing placements on high-trust domains that feed the LLM databases. Look for unlinked brand mentions; getting talked about is sometimes just as important as the backlink itself in an LLM world.
4. Target the "Messy Middle" and Transactional Queries
Users still fiercely prefer human-verified websites when money, health, or significant decisions are on the line (Your Money or Your Life - YMYL).
While you might lose traffic for basic definitions, you can capture highly qualified leads by targeting mid-to-bottom funnel keywords. Focus your SEO efforts on commercial intent: comparison guides, hyper-specific use cases, and deep-dive tutorials that an AI Overview simply cannot comprehensively cover without risking hallucination.
The Bottom Line
Are AI Overviews stealing clicks? Yes, for top-level informational queries. But the traffic you are losing is likely traffic that was never going to convert anyway.
The businesses that succeed in the next decade will be those that stop fighting the AI and start feeding it. By embracing Large Language Model Optimisation (LLMO), structuring your data efficiently, and leveraging true human experience, you can ensure that when a potential client asks ChatGPT for the best solution in your industry, the AI recommends you.