Anthropic has unveiled two significant updates today: a new Research feature and integration with Google Workspace.
With this announcement, the AI assistant Claude can now establish connections to Gmail, Google Calendar, and Google Documents, effectively becoming what the company calls “your frontline workplace assistant.” Users can leverage Claude to streamline workflows by allowing it to retrieve crucial information from various sources. For instance, it can generate meeting notes, highlight action items from follow-up email conversations, and search for contextually relevant documents. This positions Claude in direct competition with Microsoft’s Copilot and other similar workplace AI solutions.
In addition, the new Research feature is designed to enhance how Claude searches and interprets information. This innovation follows comparable launches from both OpenAI and Google, which introduced tools named Deep Research a few months earlier. Rather than operating through traditional single-query AI searches, Claude now functions “agentically” — responding to user queries by conducting multiple, interconnected searches. These results are accompanied by citations, enabling users to verify the accuracy of the information presented.
Scott White, a product executive at Anthropic, shared with Technology News that the company aims for a “sweet spot” of one to five minutes per query within the Research feature. He stated, “You have a problem that you want to solve, and you can work with Claude directly to solve it without really breaking your workflow.”
The Research ability is currently in early beta for Max, Team, and Enterprise plans available in the United States, Japan, and Brazil, while the Google Workspace integration is in beta for all paid users. However, for Team and Enterprise accounts, administrators will need to activate Google Workspace access before individual users can connect their accounts. When queried about the possibility of extending the Research feature to lower-tier subscriptions, White expressed optimism about future availability, indicating it is expected to reach the $20 monthly Pro tier soon.
For Enterprise clients, Anthropic is also introducing “Google Drive cataloging,” which employs retrieval augmented generation (RAG) methods to enhance document search capabilities. This feature autonomously scours an entire document library to ensure that important information can be accessed, even if it is buried in outdated files.
There are two notable concerns surrounding a chatbot’s capability to retrieve information both from the internet and within an organization: hallucinations and privacy. Given the probabilistic nature of these systems, there is potential for them to misinterpret complex inputs or generate inaccurate information. White advised users to check citations and verify details to ensure accuracy in the information presented.
On the subject of privacy, there have been concerns about AI agents being susceptible to prompt injection attacks, which could potentially expose sensitive information. White refrained from discussing specific safety measures but assured that rigorous user-level authentication is maintained, ensuring that Claude only accesses documents that users have authorized through their login credentials.
In its press release, Anthropic emphasized that these updates mark the beginning of a new era for working with Claude, with plans to broaden the available contexts and deepen the reporting features in the near future.
