Google Agentspace - AI for Everyday Work

Google Agentspace, the latest innovation from Google Cloud, aims for an ambitious goal: to redesign the future of work and significantly increase productivity in companies. Just imagine if you could really find what you're looking for in the company intranet or business apps and automate your daily tasks with AI.

Powerful Combination of Search, Generative AI, and Agents.

At its core, Agentspace brings together the advanced AI models of Gemini, the proven Google (Enterprise) Search, and all your company data, no matter where it resides, in a unified platform. The result? Effective access to various information and media, as well as the ability to automate daily tasks through personalized assistants or specialized agents.

Agentspace offers you a Google-level search across all connected corporate data, which, thanks to multimodal AI, can understand not only text but also images, videos, and soon audio streams. The platform uses and creates knowledge graphs from your data to improve search accuracy and personalize search results. You can chat directly with your data, have it summarized, or use it as a source for creative tasks. Agentspace includes other Google AI innovations such as NotebookLM Enterprise, Imagen, and Veo, securely and effectively integrated.

To make this work, Agentspace comes with a large and growing number of enterprise connectors to connect to your data. These pre-built connectors help integrating with common systems like Jira, SharePoint, Outlook, Salesforce, Slack, and Google Workspace, etc. Over 100 enterprise applications are already released or planned.

Users only see the data they are authorized to access. The access rights of the connected data sources are respected via Access Control Lists (ACLs). The connectors are designed to be as easy as possible for an administrator to set up and have minimal impact on existing systems.

Agents: Growing Ecosystem of Digital Colleagues

In Google Agentspace, you will find predefined Google AI agents such as the Deep Research Agent or the Idea Generation Agent, and you can create your own custom agents. This is relatively easy via a chat interface with the Agent Designer for no-code. For developers, there are various options with Vertex AI Agent Builder or Agent Development Kit and Agent2Agent Protocol to build your own agents and even have them collaborate in multi-agent systems. Administrators can centrally deploy and manage the agents. Additionally, Agentspace is planned to have a marketplace in Google Cloud for the integration of third-party agents.

Your Data, Your Rules

The platform uses the Google Cloud infrastructure and leverages known controls regarding security, data protection, and compliance.

Your data: Strict separation of customer data and no training of AI models or product improvements with your data. Also: process your data exclusively in the chosen region and protect it additionally with, for example, VPC Service Controls (data residency).

Your keys: Just like almost all services in the Google Cloud, Agentspace offers data encryption, both at rest (encryption at rest) and in transit (encryption in transit). If you need or want, you can use your own keys (customer-managed encryption keys, CMEK). 

Your control: using audit logging, all accesses and activities can be logged (Access Transparency). This way, you can keep track of the company-compliant use of your Agentspace.

The Future of Work is Intelligent and Connected. 

Google Agentspace is offered as a fully managed SaaS application, making it straightforward to get started. After setup, Agentspace is accessible to all employees and immediately usable. My recommendation: start with a well-defined pilot. Which use case has the greatest potential for the company? Which 2-3 data sources should be connected initially? How can the success of the pilot be measured?

Ready to Get Started?

What do you think? Feel free to share your thoughts in the comments or contact me for an individual demo to experience Agentspace in action.