Google adds new features in Search as AI race intensifies

Google has broadened its AI search strategy by incorporating Gemini 2.5 Pro and Deep Search into its long-dominant Google Search platform. This action occurs as it encounters strong competition from OpenAI, which unveiled a ChatGPT agent on Thursday. The tech giant announced on July 16 that Google AI Pro and AI Ultra subscribers can now choose the Gemini 2.5 Pro multimodal model from a drop-down menu located in the AI Mode tab in Search. Subscribers can now utilize Deep Search, Google’s sophisticated tool for thorough research.

Google is introducing a new feature in Search: AI-driven calls to select local businesses. Users have the option to request Search to contact businesses for pricing and availability details. The feature for business calling is now accessible to all users of Search. A voice created by the Google Duplex system notifies businesses that an AI is making a call, and users get the details via text. According to Google, humans will step in during the process when needed.

The updated features in AI Mode and Search represent Google’s ongoing enhancement of its search approach as generative AI (GenAI) leader OpenAI and AI search company Perplexity aim to challenge the cloud provider’s supremacy. They also enter a market where users are increasingly relying on large language models (LLMs) like OpenAI’s ChatGPT for answers instead of primarily using Google for searching.

The day following the Google announcement, OpenAI introduced the ChatGPT agent, allowing users to request ChatGPT to handle tasks similar to a personal assistant, including planning and purchasing breakfast ingredients, analyzing competitors, making reservations, or designing a slide deck. ChatGPT features a visual browser that navigates the web using a graphical interface along with a text-oriented browser for straightforward, reasoning-focused web inquiries.

Users can link it to Gmail, GitHub, or even redirect it to access any site, enabling it to explore more extensively and comprehensively. If the OpenAI agent functions as claimed, it would tap into some of the promise that Apple, for instance, showcased when it launched Apple Intelligence a little over a year ago. Many of those agentic characteristics aren’t functioning yet, and it’s uncertain when they will become active.