Anthropic has introduced a new research capability in its Claude AI model, allowing the system to perform multi-step investigations autonomously. Claude can deliver well-reasoned responses with verifiable citations within minutes, aiming to achieve "a balance between speed and quality."
This AI startup claims that Claude's responses to prompts are "autonomous," capable of independently determining what information is needed to construct the right response. It "automatically explores different angles of your question and systematically addresses open-ended inquiries," Anthropic stated in its press release.
The research function is currently in early testing, accessible to Claude Max, team, or enterprise plan users in the US, Japan, and Brazil. Once activated in account settings, it can be used directly within the Claude interface.
Anthropic's announcement follows similar updates from OpenAI, Google, and Microsoft this year, as these companies also add autonomous research capabilities to their proprietary chatbots. Notably: Anthropic released a study this month showing that its Claude model and DeepSeek's R1 sometimes fabricate explanations for their internal reasoning processes.
Claude's New Google Integration
Alongside the research function release, Anthropic has launched two new Google integrations for Claude: the AI can now connect to users' Gmail and Google Calendar to provide more useful, context-aware responses. These additions build on the Google Docs integration first introduced in November last year.
According to Anthropic, once Google Workspace access is enabled, Claude can find hidden meeting notes or action items within emails and calendar events. This eliminates the need for users to manually upload files before interacting with the chatbot.
These integrations can even work alongside the research function. For instance, it can alert parents if commitments in their calendar conflict with activities listed on their child's school website, or if severe weather is forecasted. Like the research function, Claude provides citations when using Google Workspace integrations, clearly showing the sources of information used in its responses.
Google Workspace integrations are currently available only in testing form to paying Anthropic users, accessible through their profile settings. If part of a team or enterprise plan, administrators must first enable Google Workspace access across the entire domain before individual users can connect their accounts.
Enhanced Generation Through Document Search
Anthropic has also rolled out a Google Docs cataloging function for enterprise administrators, utilizing retrieval-augmented generation technology. This essentially allows it to search for relevant information across entire document ecosystems without requiring users to specify exact file locations.
As a result, even if the required information is buried deep within lengthy or forgotten documents, or emerges through patterns distributed across multiple files, Claude can still retrieve it. Anthropic states that it uses "enterprise-grade security" to protect the organizational data it accesses.
Claude isn't the first AI chatbot to gain Google Workspace integration, as Gemini and ChatGPT also offer these integrations, and Microsoft's Copilot can access Microsoft 365 applications like Word, Excel, and Outlook. Despite trailing behind ChatGPT in total user numbers, Anthropic positions Claude as a competitive enterprise tool by expanding its functionality and integration capabilities.
