ChatGPT’s Main Competitor Learns to Do Deep Research and Dig into Gmail

Anthropic, the developer of the AI ​​assistant Claude, has unveiled a new tool called Research for deep research using AI. The company also announced integration with Google Workspace apps: the AI ​​assistant will be able to directly connect to Gmail, Google Calendar, and Google Docs.

Image source: Anthropic

Instead of manually searching through a huge number of emails, invitations, and documents, Claude will do it for you. The AI ​​can find information for a given period, highlight key actions from email threads, and offer additional context no worse than Microsoft Copilot, The Verge notes.

In addition, Anthropic has launched a new feature called Research, which, according to its developers, fundamentally changes the approach to searching and analyzing information. Unlike regular AI search, Claude now works as an AI agent, performing a series of interconnected queries, forming a more accurate and informed search result, and each answer is accompanied by links to sources.

Image source: Anthropic

The Research feature is already available in early beta for Max, Team, and Enterprise users in the US, Japan, and Brazil. The Google Workspace integration is in beta and is available to all paid users, but must be enabled by an administrator for enterprise accounts. Research is planned to be released soon for Pro subscribers ($20 per month).

For corporate clients, Google Drive cataloging is also available, using RAG (research engine optimization, including external search) methods to find information even in long-forgotten documents. Claude will be able to automatically scan the entire company database for the data you need.

However, as the model’s capabilities expand, the question of accuracy and security arises. Like other AIs, Claude can misinterpret data or provide fictitious information. “We encourage users to always check sources, read them, and ensure the accuracy of the information,” Anthropic emphasizes. As for privacy, the company assured that its AI assistant only works with documents that the user has granted access to as part of their authorization.

Anthropic is also reportedly launching a voice AI that you can talk to. The new feature, called Voice mode, will be available in the chatbot this month and is likely to compete with a similar voice mode in ChatGPT. According to Bloomberg, users will have three voice options in English: Airy, Mellow, and Buttery. The development of this feature was previously mentioned by Anthropic’s head of products, Mike Krieger, who confirmed that the company is testing prototypes. References to the voice mode were also found in the code of the iOS app Claude.

The company said this is just the beginning of a series of updates that “reimagine how you interact with Claude.” In the coming weeks, it expects to further expand the available context and improve the AI ​​model’s analytics capabilities.