The world of artificial intelligence is crossing the boundary between "talking machines" and "working machines." While in previous years we learned how to use LLMs (large language models) to generate text or code, we are now entering the era of agentic AI. This means that AI will no longer just answer questions, but will be able to plan, use tools, and complete tasks for us. The collaboration between Deloitte and Google Cloud aims to accelerate this shift on a global scale.
What is "Agentic AI" and why is it a fundamental shift?
To understand the significance of this announcement, it is necessary to define what an agent is in the context of AI. A traditional chatbot (like the basic version of ChatGPT or Gemini) works on a question-and-answer principle. If you tell it: "Plan me a trip to London," the chatbot will write you an itinerary. An AI agent, however, works differently. It can connect to your calendar, search for the cheapest flights, book a hotel, and subsequently send you a confirmation email.
Deloitte is focusing on exactly this. Using its Deloitte Ascend platform and Google Cloud technologies, it is creating systems that can autonomously solve complex tasks. The key element here is Google's Agent2Agent protocol, which allows different AI agents (e.g., an agent from Salesforce and an agent from Google Workspace) to communicate with each other and collaborate on a single goal. This is a technological milestone that enables the connection of isolated software ecosystems.
Scaling with Gemini Enterprise
At the heart of this transformation is the Gemini Enterprise model. Google Cloud is investing enormous resources into this ecosystem – as part of innovations for partners, a fund of $750 million was announced to support the development of these very agentic solutions.
Deloitte is already using this technology internally. Currently, more than 25,000 of its employees use it, with plans to expand the license to 100,000 people. Examples of practical use include internal tools like Scout (a personal learning assistant) or specialized orchestrations for marketing teams.
Technology comparison: Gemini vs. competition
When evaluating the capabilities of agentic AI, it is important to compare Gemini with the main competitors on the market:
- Google Gemini Enterprise: Excels thanks to deep integration into Google Workspace (Docs, Gmail, Sheets). For companies that already run on Google Cloud, agent implementation is the smoothest due to direct access to data and infrastructure.
- OpenAI (GPT-4o): Currently leads in the area of pure logic and the broad "GPTs" ecosystem, but its integration into enterprise workflows requires more external tools than Google's.
- Anthropic (Claude 3.5 Sonnet): Highly valued for its ability to finely understand context and a "human" writing style, which is great for creative agents, but in complex cloud service integration it still lags behind Google.
Impact on Czech and European business
For the Czech market and European companies, this announcement has two main aspects: availability and regulation.
Availability: Google Cloud and Gemini services are fully available in the Czech Republic. For Czech companies that use Google Workspace, Gemini Enterprise represents a direct path to automating administration, customer support, or data analysis. Since Google invests in local data centers and model optimization, language support for Czech is at a very high level with these models, which is key for successful agent deployment in our region.
Regulation and security: In the European Union, a key factor is the EU AI Act. Deloitte, within its cooperation with Google, emphasizes the Trustworthy AI framework. This means that agents must be transparent, their decision-making processes must be auditable, and they must strictly comply with GDPR rules. For European companies, the ability of Deloitte to implement AI that is in compliance with European legislation is a fundamental argument for adoption.
Price and implementation
Since this is a solution for the Enterprise segment, there is no publicly disclosed fixed price for individuals. Gemini Enterprise is usually provided as part of Google Workspace or Google Cloud under B2B licensing models for companies. Prices depend on the number of users and the scope of cloud services used. However, for small and medium-sized businesses in the Czech Republic, it is possible to start with lower tiers of Google Cloud, which allow testing of basic AI functions at much lower costs than full implementation by Deloitte.
According to the Deloitte 2026 State of AI in the Enterprise report, approximately 60% of surveyed organizations already use AI tools in daily work. This confirms that we have long since moved from the experimental phase to the phase of massive deployment in the real world. The original report on this announcement can be found here.
What is the main difference between a regular AI chatbot and an AI agent?
A chatbot responds to text queries and generates answers. An AI agent has the ability to "act" – it can plan steps, use external software, connect to other systems, and complete entire processes (e.g., purchasing, booking, data analysis) without constant human prompting.
Is Gemini Enterprise safe for sensitive corporate data within the EU?
Yes, the Google Cloud Enterprise version is designed with an emphasis on security and privacy. Data used within Gemini Enterprise is not standardly used to train Google's public models, which is fundamental for GDPR compliance and protection of trade secrets within the EU.
Can a Czech small business use these technologies, or is it only for corporations?
While complex services from Deloitte are intended for large corporations, the Gemini technology itself and automation possibilities in Google Workspace are also available to smaller companies in the Czech Republic. Smaller businesses can start with automating simple tasks (e.g., sorting emails or generating reports) using standard Google Cloud subscriptions.