Search engines are no longer just lists of links. In 2025, AI‑powered answers, chatbots, and voice assistants are redefining how people find information. Google’s AI Overviews (AIOs) roll generative answers into the search results page, summarizing information and linking to sources. Microsoft has integrated Copilot into Bing, and AI chatbots such as ChatGPT, Perplexity, and Claude have become stand‑alone information gateways. The rise of generative AI in search is causing a seismic shift in organic traffic and requires marketers to adapt their strategies.

This long‑form guide explores the latest data on AI‑powered search, explains why Generative Engine Optimization (GEO) has become essential, and shows how voice interfaces are merging with generative search. It closes with action steps that digital marketers can take to protect visibility and remain competitive.

The explosion of AI Overviews and generative search

How AI Overviews are impacting search behavior

Studies in mid‑2025 reveal that AI Overviews are altering click‑through behavior. WordStream’s analysis of thousands of search results found that:

AI Overviews appear in around 55 % of Google searches, and the feature’s presence grew 115 % between March 2025 and May 2025.
The feature crowds out traditional organic results — AIOs take up 42 % of the desktop screen and 48 % of the mobile screen.
70 % of consumers trust generative AI search results, and 79 % expect to use AI‑enhanced search within the next year.
When an AI Overview appears, click‑through rates (CTR) drop by ~35 % on average. Some sites saw 20–40 % declines in traffic, and organic CTR is predicted to fall another 25 % by 2026.

Fresh research from Omnius’s AI Search Industry Report 2025 paints an even starker picture. It collates data from multiple SEO tools and shows that:

Ahrefs measured a 34.5 % reduction in position‑1 CTR when AIOs appear, while Amsive found an average 15.49 % decline.
BrightEdge’s data suggests search impressions rose 49 % year‑over‑year, yet click‑through rates fell 30 %.
Across several studies from January 2024 to May 2025, CTRs dropped 37–40 % when AIOs were present.
AIO coverage is expanding rapidly — the report notes AIOs were triggered for 6.49 % of queries in January 2025, 7.64 % in February and 13.14 % by March, representing 72 % month‑over‑month growth.
ChatGPT holds ≈59 % of the generative chatbot market, while Microsoft Copilot and Google Gemini share 13–14 %.

These numbers underscore a fundamental shift: ranking first in traditional blue‑link SEO is no longer enough. Users increasingly read AI summaries and never click through to websites. Google still dominates with ~89.6 % of global queries, but AI summaries mean less traffic from each query. Marketers must therefore adjust their optimization strategies.

Voice search and conversational interfaces

The fusion of voice assistants and generative AI is accelerating. By mid‑2024 there were about 149.8 million voice assistant users in the United States, a number projected to reach 153.5 million in 2025. The latest voice search statistics show:

About 20.5 % of the global internet population uses voice search, amounting to over a fifth of all users.
8.4 billion voice assistants are estimated to be in use globally (many households have multiple devices).
In the U.S., Siri has ~84 million users, Google Assistant ~88.8 million, and Amazon Alexa ~75 million.
Voice assistants answer 93.7 % of queries accurately, and there are over 1 billion voice searches per month.
76 % of consumers use voice search to find local businesses.

As generative search becomes multimodal, these voice interfaces will likely rely on the same large language models (LLMs) powering AIOs. Marketers must therefore optimize content for voice queries, which are longer and more conversational, and ensure local SEO information (such as NAP data) is accurate.

From SEO to GEO: optimizing for generative engines

Generative engines assemble answers by querying large amounts of content, citing sources when helpful. Generative Engine Optimization (GEO) expands traditional SEO by ensuring your content is favored by these AI models. First Page Sage explains that GEO builds on SEO but prioritizes factors that influence LLMs: authoritative data, topical expertise and trust signals. Their research shows that being included in high‑authority directories and academic research influences generative engine recommendations by 55 % for ChatGPT and 45 % for Gemini, while mid‑tier sources such as mainstream media influence around 25–30 %.

Key GEO best practices include:

Build deep topical authority. Organize your website into thematic pillars and cluster supporting content around them. Generative engines weigh search intent heavily, so content must answer questions comprehensively and logically.
Secure high‑authority citations. Appearing in research papers, data‑driven reports and widely cited articles increases the likelihood that LLMs will extract information from your pages. Positive reviews and awards also influence generative recommendations.
Provide structured data and transparency. Mark up content with schema for articles, FAQs and how‑tos so AI can easily parse the information. Offering clear citations, publishing methodology and showing dates helps AI validate the reliability of your content.
Leverage first‑party and user‑generated data. Generative models often pull from topically rich Q&A threads, community forums and user reviews. Encouraging authentic user contributions not only builds trust but also increases the chances of being cited.
Optimize for conversational queries. Use natural language that mirrors the way people ask questions aloud. Focus on succinct answers to “how,” “why” and “what” queries, and include synonyms and long‑tail phrases.
Monitor citations and refine. Tools like Perplexity and Copilot show their sources. Keep track of which pages are cited, analyze why they were chosen and refine content accordingly.

The convergence of AI and voice search

Generative search is converging with voice search to create a more human, multimodal experience. Google’s SGE accepts image inputs and performs live camera queries; it integrates voice, images and text. Voice assistants on phones and smart speakers now use advanced language models to converse naturally. This convergence has several implications:

Semantic depth matters. Voice queries are longer and phrased as full questions. Content optimized for natural language will perform better in both voice search and AIO contexts.
Local intent is critical. With 76 % of consumers using voice search to find local businesses, local SEO — accurate Google Business profiles, local keywords and customer reviews — becomes even more important.
Speed and concise answers. Voice results need to be delivered quickly. Ensure pages load fast and provide succinct answers near the top.
Screen‑free experiences. As voice search grows, some sessions will be completely screenless. Brands must design “spoken UX,” where content is structured to be read aloud or summarized.

Action plan for marketers

Audit your content for AI readability. Evaluate whether your pages provide clear, authoritative answers that AI can parse. Check if you have high‑quality outbound and inbound links.
Invest in thought leadership and data‑driven research. Publish original studies, whitepapers and surveys. These are exactly the types of sources generative engines prioritize.
Encourage and curate user reviews. Positive reviews and awards influence generative engines, so proactively encourage satisfied customers to leave feedback.
Implement structured data. Use schema markup for products, FAQs, how‑tos, events and local business data. This helps AI models understand your content and increases the chance of being cited.
Optimize for conversational queries. Research long‑tail, question‑based keywords and build content that answers them clearly. Provide concise definitions and summaries near the top of your page for voice search.
Monitor generative citations. Use tools like ChatGPT, Bing Copilot and Perplexity to search your target keywords. Note which sources appear in the AI response and refine your strategy accordingly.

Conclusion

Generative AI search and voice interfaces represent the most dramatic change to search since Google launched. AI Overviews are already present in more than half of search results and are expanding rapidly. The shift is reducing traffic to traditional organic results, but it also opens new visibility channels. By understanding how generative models select and cite content and by optimizing for voice and conversational queries, marketers can stay ahead. Implementing Generative Engine Optimization, investing in authoritative content, and embracing first‑party data will ensure your brand remains discoverable in this AI‑driven search landscape. If your business needs help navigating this new AI world, Orange SEO has AI-Optimized SEO services that can help you succeed online.

FAQ: Generative AI Search, AI Overviews, and GEO

1) What is Generative Engine Optimization (GEO)?
GEO is the practice of tailoring content so generative search systems (e.g., AI Overviews, Copilot, Perplexity) confidently select, summarize, and cite your pages. It extends SEO with more emphasis on topical authority, structured evidence, freshness, clarity, and trust signals.

2) How is GEO different from traditional SEO?
Classic SEO targets blue-link rankings and snippets. GEO targets answer inclusion inside AI summaries. That means deeper coverage of questions, structured data, transparent sources/methods, consistent terminology, and concise, quotable passages.

3) Does optimizing for GEO hurt my traditional SEO?
No. Most GEO tactics—expert depth, schema, internal linking, first-party data—also improve organic rankings and engagement. Think “SEO++,” not “SEO vs. GEO.”

4) What content formats perform best in AI Overviews?
Comprehensive explainers, step-by-step how-tos, comparison matrices, FAQs, data-backed guides, and original research. Pages that answer the query and adjacent follow-ups in one place are favored.

5) How do I make my content more “quotable” for AI answers?
Use clear H2/H3 question headings, tight 2–4 sentence answers near the top, bullet lists for steps, and definitive definitions. Include key stats or formulas with brief context so a model can lift them cleanly.

Sources

[1] wordstream.com
[2] wordstream.com
[3] omnius.so
[4] omnius.so
[5] omnius.so
[6] omnius.so
[7] omnius.so
[8] omnius.so
[9] omnius.so
[10] yaguara.co
[11] yaguara.co
[12] yaguara.co
[13] yaguara.co
[14] yaguara.co
[15] firstpagesage.com
[16] firstpagesage.com
[17] firstpagesage.com
[18] omnius.so
[19] omnius.so