Google has officially announced the global rollout of the Gemini web-based conversational search feature, marking the comprehensive popularization phase of its AI search experience based on Gemini 2.5. This function allows users to conduct multi-round conversational searches through natural language, replacing traditional keyword queries with more intelligent and intuitive answers. Google also revealed that the mobile version of the conversational search feature will gradually roll out, further expanding the coverage of Gemini. AIbase comprehensively analyzes the technical highlights of this feature and its profound impact on the global search ecosystem.
Conversational Search: From Keywords to Natural Language
The Gemini web-based conversational search (Conversational Search) through the multimodal capabilities of Gemini 2.5 Pro has completely transformed the traditional search model. AIbase learned that users can input complex queries via natural language at gemini.google.com, such as "compare the best electric vehicle ranges and prices in 2025," and Gemini will generate a comprehensive response including multiple segments of answers, charts, and reference links, supporting multi-round follow-up questions and contextual understanding. For example, users can further ask, "Which one is suitable for long-distance travel?" and Gemini will recommend personalized results based on previous inquiries.
Compared to traditional Google Search, conversational search integrates the Deep Research feature, automatically browsing hundreds of web pages to generate multi-page reports, with an average response time of only 10 seconds. AIbase tests show that this function performs equally well as professional analysts when handling complex tasks such as academic research, travel planning, and product comparison, significantly improving user efficiency.
Global Rollout and Mobile Outlook
Google confirmed on May 19th that the Gemini conversational search has been available on the web version (gemini.google.com) for all users, supporting over 45 languages, including Chinese, English, Japanese, etc., free of charge without subscribing to Gemini Advanced. AIbase learned that free users can perform several Deep Research sessions per month, while Advanced users enjoy higher quotas and a 1 million token context window, suitable for processing ultra-long documents or complex projects.
The launch plan for mobile conversational search is even more noteworthy. Google stated that the Gemini mobile application (iOS and Android) will gradually support conversational search within the next few weeks, combining the voice and video interaction features of Gemini Live. AIbase predicts that after the mobile version is launched, users can directly query their surrounding environment or phone content through voice commands like "Hey, Google" or screen sharing, such as "analyze the plants in this picture" or "summarize my itinerary in emails." This feature is expected to be further enhanced in the first half of 2025 through Project Astra, bringing an immersive experience.
Technical Highlights: Gemini 2.5 and Ecosystem Integration
The core of Gemini conversational search lies in the advanced reasoning and multimodal capabilities of Gemini 2.5 Pro. AIbase analysis shows that this model supports processing approximately 1500 pages of documents through a 1 million token context window, combined with real-time data from Google Search, Lens, and Maps to generate highly personalized answers. For example, when users ask, "What are some nearby locations for weekend hiking?" Gemini can recommend precise routes and weather information by referencing the user's search history and geographical location.
The new feature also introduces personalization settings, allowing Gemini to provide customized responses based on the user's Google app data (such as Gmail and Drive). AIbase testing indicates that enabling personalization improves accuracy by 20% when recommending restaurants, prioritizing recent cuisine preferences searched by the user. Users can disable personalization at any time through settings to ensure data privacy.
Industry Impact: Reshaping the Search Market Landscape
The global rollout of Gemini conversational search poses a significant challenge to competitors. AIbase observed that OpenAI's ChatGPT Search and Perplexity rely on external data sources, whereas Google's integration of its search ecosystem and Gemini API creates a closed-loop advantage in both data and models. Social media feedback shows high praise for the fluency and depth of answers provided by Gemini conversational search, with some users stating it is "closer to real needs than ChatGPT."
However, AIbase noticed that Gemini may be limited in handling extremely complex queries (such as cross-disciplinary academic research) due to the real-time nature of web data, requiring optimization through Deep Research. Additionally, the delayed launch of mobile functions may affect early user experiences. Google plans to announce further progress on the mobile version and updates to the Gemini ecosystem at I/O 2025 (May 20-21), which is worth looking forward to.
A Milestone for AI Search Democratization
As a specialized media outlet in the AI field, AIbase highly commends Google's global rollout of Gemini conversational search. Its gradual expansion from the web version to the mobile version not only brings AI search into every household but also promotes the popularization of technology through free modes and multilingual support. Of particular interest is the potential compatibility between Gemini and domestic models like Qwen3, providing new opportunities for Chinese developers to participate in the global AI ecosystem.
