Google's AI Mode can now find restaurant reservations for you - how it works

SvetaZi/iStock/Getty Images Plus via Getty Images

ZDNET's key takeaways

• Google's AI Mode can now suggest restaurant reservations.
• Users can also now receive personalized dining recommendations.
• Google is aiming to build a more dynamic and personalized web.

AI Mode can now find restaurant reservations, Google announced in a blog post Thursday. Users can now request reservations in local restaurants according to constraints like party size and dietary restrictions, and the system will automatically respond with available options.

Also: I used Perplexity to make a restaurant reservation

Google upgraded AI Mode in search with more agency in an effort to make the feature more autonomous and personalized to the unique preferences of individual users. The idea is to deliver a smarter search engine that can actually take useful action on behalf of users, and read between the lines of prompts to provide helpful information even when users aren't precisely sure what they're looking for.

So far, those agentic actions are pretty mundane, but they hint toward a not-so-distant future in which the online search experience feels more like a two-way conversation between human users and AI.

"AI Mode does the legwork and links you directly to the booking page, so you can easily take the last step and finalize your reservation," Google writes in its blog post. The feature leverages Project Mariner, Google's web-browsing AI agent, and was developed in partnership with OpenTable, Resy, SeatGeek, Ticketmaster, and other booking services. It's being rolled out now to Google AI Ultra subscribers in the U.S.

More personalized upgrades 

The company also announced that users in the US who have opted into AI Mode in Search Labs can now receive tailored dining recommendations based on their search histories. For example, if you were to prompt the system with: "I'm meeting a colleague for coffee, please suggest a good spot nearby," it might scour your previous searches to deduce that you're particularly partial to cafés that offer iced oatmilk lattes, reliable Wi-Fi, and plenty of electrical outlets, and suggest someplace that meets all of those criteria.

Also: Google reveals how much energy a Gemini query uses - in industry first

Users can control how much of their previous search history and personal information they want to be used to inform tailored recommendations in settings, Google wrote in its blog post.

Additionally, Google debuted a sharing feature in AI Mode for US users, allowing them to invite others to view and interact with the responses generated by the system. 

"This can be particularly helpful for tasks you want to collaborate on, like planning a trip or a birthday party," Google writes.

A more agentic web

The new features are part of a broader effort from Google and other tech developers to build a more dynamic and personalized internet with the help of AI.

Also: Scammers have infiltrated Google's AI responses - how to spot them

Historically, navigating the web has been a mostly one-way experience: human users have had to carefully specify requests to a search engine, say, and then search through a long list of links to (maybe) find what they were looking for. But now Google and other companies are debuting new features engineered to make the online experience more of a dynamic and collaborative interaction between human users and machines, in which the latter is better able to take action without being specifically prompted to do so, and personalize their outputs to individual users. (This personalization aspect has become a critical focus within OpenAI following its rocky release of GPT-5.)

Google's AI Overviews feature, for example, which automatically generates responses to user queries, was launched in response to the viral success of ChatGPT and the conversational style of browsing that the chatbot introduced. 

Also: How to get rid of AI Overviews in Google Search: 4 easy ways

More recently, Google also upgraded its Flights tool so that users can simply describe in natural language a rough approximation of the trip they'd like to embark upon, and the system will fill in the blanks by finding destinations that fit that description, along with affordable flight plans.