Humanity announced on Tuesday that Claude, an AI chatbot, has been integrated with Google Workspace, allowing you to search and browse emails in Gmail.
The integration is first rolled out in beta to subscribers on Anthropic’s Max, Team, Enterprise and Pro plan. Administrators who manage multi-user accounts must enable integration before users can connect Google Workspace and Claude accounts, according to humanity.
Google Deepmind’s Gemini Chatbot is also integrated with Workspace, while Openai’s ChatGPT is integrated with Google Drive. However, Anthropic is one of the first third-party AI companies to provide a way to connect closely to Google’s productivity suite.
Humanity’s team-up with Google aims to provide personalized responses by Claude without requiring users to upload files repeatedly or create detailed prompts. Openai and Google have tried to achieve the same effect via various approaches, including adding memory features that allow chatbots to refer to past conversations in replies.
In a press release, Anthropic says that the new integration of Claude will help users organize their professional and personal lives. For example, humanity claims that this feature can help parents by searching for “weather forecasts that may affect the latest school calendars, local community events, and family planning, while scanning email and calendar events to highlight important commitments.”
Claude provides inline quotes when referring to workspace content, indicating the user exactly where certain information occurred, says humanity.
The integration does not allow Claude to provide the ability to schedule calendar events or send emails, but it could cause security concerns among some users. It is unclear how extensively Claude searches a person’s Google workspace, or whether users should instruct Claude to view a specific email or calendar event depending on the nature of the request. It is also unclear whether users can ask Claude not to search for certain sensitive emails or files.
In response to the above, a humanity spokesman told TechCrunch that the company does not train a model of user data by default and implements “strict authentication and access control mechanisms” in external services such as Workspace.
“Connections to each user or organization’s external services (such as Google Drive, Gmail) have been properly authenticated and approved for only specific users or organizations,” the spokesperson said in a statement. “Claude does not have the ability to access or transfer data between connection services for different users, as each connection is bound to a specific authentication credential for an individual user or organization.”
Humanity also announced on Tuesday the launch of a new feature that will conduct multiple web searches to generate detailed answers. Claude Research is positioned as a competitor of Openai and Google’s “deep search” agents, offering the “optimal trade-off” between speed and inclusivity, says humanity.
According to a human spokesman, Claude’s research usually takes less than a minute to compile the information. It’s faster than some of its rival deep research agents. However, Claude Research does not use custom models and instead takes advantage of Claude’s recently launched web search capabilities.
The company deploys Claude Research to MAX, teams and enterprise plan subscribers in the US, Japan and Brazil. According to humanity, it will soon come to professional customers.
These updates are part of humanity’s broad efforts to attract users to AI subscription plans that have the capabilities to make Claude more capable and convenient. Claude is gaining popularity, but data compiled by Simarweb shows that it reached 3.3 million web users in March, but Anthropic’s user base remains ChatGpt’s user base.
Source link
