Google is letting users test a new artificial intelligence tool that connects information from various apps like Gmail and Google Photos to provide personalized answers in its chatbot.

The feature, called Personal Intelligence, is available for personal accounts, Google said in a blog post on Wednesday.

“Whether it’s connecting a thread in your emails to a video you watched or finding nuance in your photo library, Gemini now understands context without being told where to look,” Josh Woodward, vice president of Google Labs and Gemini app, wrote in the post.

It’s Google’s latest attempt to build more reasoning AI capabilities within it Gemini app and lure users as the company goes head-to-head with OpenAI and others in the increasingly competitive generative AI market. While Gemini could already retrieve information from Google apps, the company said on Wednesday Gemini 3 can now “reason across your data to surface proactive insights.”

The new feature will compete with Apple Intelligence, Apple’s personal AI system that integrates apps to help with writing, image creation, and understanding context. Apple said Monday it chose Google to power its AI features, including a major Siri upgrade expected later this year. 

Personal Intelligence will first be available to Google AI Pro and AI Ultra subscribers in the U.S., the company said, and will be added to Google’s search tool “AI Mode” in the future. The feature will be off by default, Google said.

Woodward, who has launched successful products in Gemini in the past year, warned that the beta version of the new feature hasn’t eliminated mistakes, and is asking users to give feedback.

“Gemini may also struggle with timing or nuance, particularly regarding relationship changes, like divorces, or your various interests,” he wrote. For sensitive topics like health, “Gemini aims to avoid making proactive assumptions” but “will discuss this data with you if you ask,” Woodward added.

Woodward also said the company doesn’t train its AI models “directly on your Gmail inbox or Google Photos library” but does utilize “limited info, like specific prompts in Gemini and the model’s responses, to improve functionality over time.”

WATCH: Google unveils ‘Gemini 3 Flash’ AI model focused on speed and cost