With AgentsSpace, Google has launched a powerful new tool that makes an organization’s collective knowledge readily available and actionable for employees. Think of it as a single centralized source of truth for an organization’s data, accessible with Google-quality search and Gemini’s multi-modal reasoning and generative AI abilities. As the name suggests, AgentSpace is also a hub for users to access custom AI agents within an organization. These can help workers gather information, generate new content, and even perform actions like sending emails. Google built AgentsSpace for enterprise; it provides robust security, integrates with your identity provider to enforce RBAC for your data, and has multi-language support, enabling adoption across various sectors and regions. That’s a lot of functionality in one tool – let’s break it down piece by piece.

The entrypoint for AgentSpace is a customized Search Agent where users can ask questions, select data sources, view search suggestions, generate documents – all in one place. Users familiar with Gemini or other chat-based interfaces for LLMs will feel right at home.
We’ve created a sample Agentspace app for our org that is connected to Google Workspace products. Here you can see that Agentspace responded to our question, suggested some follow-up questions, and (not shown) provided citations for all the internal documents in Google Cloud that it used. Note that we can easily filter those documents by freshness to update the response.

One of the most exciting things about AgentSpace is that users can connect multiple first- and third-party data sources to Google’s powerful search engine and AI tools in a few clicks. Agentspace also has pre-built integrations for performing actions in first- and third-party productivity tools, including email and calendar apps. Together, these connectors unlock workflows like searching Jira and Slack, generating a report with AI, and sending the content as an email – all from the same interface. Importantly, organizations can use their existing identity provider to enforce access controls for these connections.

Besides drastically simplifying the process of integrating multiple data sources and systems, Agentspace offers some very powerful tools for answering questions with data. These include NotebookLM, Search, and Gemini.

AgentSpace is built on NotebookLM Plus, the enterprise version of Google’s popular research and writing tool. With NotebookLM, users can collect or upload information (URLs, Slides, PDFs, etc.) into a notebook and generate summaries, briefing documents, study guides, and even podcasts or audio summaries based on the content. In its current iteration, users can ask questions of the AI podcast hosts – literally “chat with your data”.

If you aren’t already a NotebookLM user, go check it out today. NotebookLM can help you learn new skills or understand a large corpus of information by summarizing large amounts of data into digestible chunks. You can use it to evaluate your presentation skills (upload videos and ask for feedback), look for patterns in complicated data sets, and generate a ‘Cliff’s notes’ version of any text. For anyone who creates content based on research or source information, this is an incredibly useful tool.

Think of Agentspace as a customized, access-controlled entrypoint for your organization’s data – a single source of truth, accessible via a conversational UI built on Google’s search and integrated with Gemini’s multi-modal and reasoning abilities. It’s search, but better, smarter, and customized to your organization. Agentspace can provide a tailored experience for each worker based on permissions, search history, role, and recommendations. From a single entrypoint, users can ask questions, retrieve data summaries, identify documents to add to a NotebookLM instance, generate images or text documents, and access custom agents.

To get started with Agentspace Enterprise, you first create an app and connect it to at least one data store. Once a data store is linked to an app via Google’s pre-built connectors, users can use the data to generate search results, utilize AI tools, and take actions. Along with first-party sources, users can connect applications like Confluence, Jira, Salesforce, Sharepoint Online, and Slack. While Agentspace supports both structured (think BigQuery, JSON) and unstructured data, it is particularly adept at handling unstructured data, like HTML, PDF, image, and video. To leverage this, simply upload a corpus of documents to a Cloud Storage bucket and share it with Agentspace. Users can then search, integrate, and generate recommendations or summaries from those documents. Documents returned in search results also can be added to NotebookLM, where users can generate reports, study guides, and even chat with AI-generated podcasters about the data.

Agentspace can take actions in connected systems, if granted permissions. Out of the box, Agentspace offers pre-built connections for creating actions (e.g., send email, update Jira tickets) in popular productivity tools like Jira Cloud, Workday, Gmail, Google Calendar, Outlook email, and Outlook calendar. Enabling actions in other systems would require creating a custom agent with the required connection and permissions.

For any custom conversational experiences, Agentspace supports connecting to Dialogflow agents. Custom Dialogflow agents can be deterministic, fully generative, or a hybrid, and can connect to any other service. This makes it possible to build connections and workflows that extend beyond what is available out-of-the-box with Agentspace. For example, by connecting Agentspace to their internal systems with a custom agent, a financial institution could accomplish actions like resolving fraud disputes, refunding fees, marking credit cards as lost or stolen, or updating user records. Workers can access and use any of these agents directly from Agentspace.

Via Agentspace, users can leverage Google’s Research Assistant agent to tackle challenging research problems in a systematic, transparent, and well-documented manner. The Research Assistant is a great tool – we’re excited to see what Google has up their sleeves in terms of future pre-built agents.

Agentspace is a powerful tool both out of the box and deployed with custom agents. We asked Agentspace to highlight some of its enterprise use cases to help jumpstart your creativity:

• Quickly finding relevant information within internal databases and documents across different systems.
• Accessing expert knowledge on specific topics by querying an AI agent connected to company data.
• Generating summaries of complex documents or presentations.

• Assisting customer service representatives with real-time information retrieval.

• Generating personalized sales pitches based on customer data.
• Quickly finding relevant product information to address customer queries.
• Identifying potential leads through data analysis.

• Generating marketing copy, social media posts, and email campaigns.
• Analyzing market trends and identifying target audiences.
• Creating data-driven reports and presentations.

• Accessing and analyzing research papers and industry reports.
• Identifying potential research areas and generating hypotheses.
• Summarizing complex technical information.

• Automating repetitive tasks like report generation.
• Streamlining workflows and decision-making processes.
• Identifying potential bottlenecks and areas for improvement.

Google currently offers multiple tiers for Agentspace and as a launch partner, 66degrees can help you get started with Agentspace. Our Industry Blueprints for Agentspace empower organization’s with a strategic guide and comprehensive framework that prioritizes cross-departmental collaboration, alignment between Agentic AI outputs and business needs. Our extensive experience leveraging 1st and 3rd party data connectors ensures seamless integration of Agentspace with your existing technology stack.

Ready to unlock the full potential of AI agents and transform your business? Request a demo today to discover how our blueprints and Agentspace can unlock your organization’s collective brilliance.