Case study
Gemini Enterprise and GEAP
Gemini Enterprise AI Platform: AI agents for your business
Gemini Enterprise brings AI agents to every employee – with ready-made agents, your own agent builder and access to your company data. For your own deeply integrated agents, there is the Gemini Enterprise Agent Platform (GEAP). We introduce both and run them with you.
- Roll out Gemini Enterprise to all employees
- Build and run your own agents on GEAP
- Free choice of models: Gemini, Claude and more
In short
Gemini Enterprise is Google's AI platform for every employee: one app for Gemini, enterprise search and AI agents. The Gemini Enterprise Agent Platform (GEAP) is Google's platform for building and running your own agents. GMinds introduces both in companies in Germany, Austria and Switzerland.
Gemini Enterprise or GEAP – what's the difference?
Gemini Enterprise – the AI workspace for every employee
One app in which employees work with Gemini, search company knowledge and use agents. Ready-made agents, for research for example, and your own agents built without code (Agent Designer), connected to Google Workspace, Microsoft 365, Salesforce, ServiceNow and other systems. Since April 2026, the former Agentspace is also part of it (according to Google).
Gemini Enterprise Agent Platform (GEAP) – the platform for your own agents
This is where developer teams build, test, run and monitor their own agents and AI applications – with the Agent Development Kit, Agent Runtime, governance and more than 200 models in Model Garden (according to Google). Since April 2026, Agent Studio, Agent Search, RAG Engine and Vector Search are also part of GEAP (according to Google's release notes).
| Gemini Enterprise | GEAP | |
|---|---|---|
| Who it's for | All employees | Developer and data teams |
| How agents are built | Agent Designer, no code | Agent Development Kit, with code |
| Typical tasks | Research, writing, finding knowledge, simple workflows | Deeply integrated processes, your own applications, automation across systems |
| Connectivity | More than 100 ready-made connectors, such as Microsoft 365, Salesforce, ServiceNow and Slack | Any system via APIs and your own code, such as SAP |
| Billing | License per user | Usage-based |
| Interplay | Also uses agents from GEAP | Makes agents available in Gemini Enterprise |
Rule of thumb: Gemini Enterprise when many employees should become more productive. GEAP when a process needs a tailored solution. Most companies start with one and add the other.
Connectivity
Which systems Gemini Enterprise connects directly
Gemini Enterprise comes with more than 100 ready-made connectors (according to Google). They respect existing permissions: everyone only sees what they are allowed to see in the source system.*
Google Workspace
Gmail, Drive, Calendar
Microsoft 365
SharePoint, OneDrive, Outlook, Teams, Entra ID
Salesforce
Customers, opportunities, activities
ServiceNow
Tickets, incidents, knowledge articles
Slack
Channels and messages
Jira and Confluence
Projects, tickets, wiki
Box and Dropbox
File storage
GitHub
Code and documentation
HubSpot
Contacts and deals
* Availability depends on the edition; some connectors are in preview. As of November 2026. Overview: Gemini Enterprise connectors
Model Garden
Not just Gemini: free choice of models
The Agent Platform offers more than 200 models in Model Garden (according to Google). Alongside Gemini, Anthropic's Claude models are available, including Claude Opus 5.5 (according to Google's documentation), plus open models from other providers. OpenAI's models have been available in preview since October 2026 (according to OpenAI). This lets you choose the right model for every task – with the same security, billing and governance in Google Cloud.*
Gemini (Google)
Multimodal and deeply integrated into Google Workspace and Google Cloud.
Claude (Anthropic)
For example Claude Opus 5.5 for demanding agents, programming and long documents.
OpenAI
The GPT-6 models, available in preview since October 2026.
Open models
Models from other providers when cost, size or operation matter most in your project.
* As of November 2026.
Where AI agents take work off your plate today
Sales
Review tenders, prepare quotes, get ready for customer meetings.
Customer service
Pre-sort requests, suggest answers from company knowledge.
HR
Answer questions about policies, support onboarding.
IT helpdesk
Resolve standard requests, pre-qualify tickets.
Contracts and documents
Check clauses, flag deviations, create summaries.
Finding knowledge
Answers from SharePoint, Drive, CRM and other sources, with existing permissions.
Approach
How we introduce Gemini Enterprise and agents
AI projects rarely fail because of the model – they fail because of missing data access, unclear ownership and pilots with no path into operations. That's why we start with a concept and a target architecture for your AI platform, then pick a few use cases with measurable value, set up data access and bring the first agent into real everyday work.
Getting started: AI Business & Readiness Workshop
01
Concept and architecture
Target picture and architecture for your AI platform: identities and permissions, data access, security and governance rules, model choice, cost framework and how Gemini Enterprise and GEAP work together.
02
Choose use cases
Work with your business units to find the tasks with the biggest leverage and make the value measurable.
03
Connect data
Connect Workspace, Microsoft 365, SAP and other sources with permissions, so everyone only sees what they are allowed to see.
04
Introduce Gemini Enterprise
Pilot group, role-based training, first custom agents in Agent Designer.
05
Custom agents on GEAP
Where a process needs more: a tailored agent with the Agent Development Kit, tests and approvals.
06
Operations and governance
Monitoring, cost control, rules for agents and continuous improvement as a managed service.
Gemini Enterprise Agent Platform (GEAP)
Build and run your own AI agents
When the ready-made agents in Gemini Enterprise aren't enough, we build your own agents on the Gemini Enterprise Agent Platform – connected to your data and systems, securely operated and monitored. Code and configuration live in your own Google Cloud project, and the finished agent is available to your team directly in Gemini Enterprise.
Agents with the Agent Development Kit
We build your agent with the Agent Development Kit: with a clearly defined task, set tools and approvals, tested on real cases from your daily work.
Connection to ERP, CRM and your own systems
Your agent doesn't just read, it acts: via APIs it works directly in ERP, CRM and other systems, for example to create an order or update a ticket.
Answers from your verified data
We connect your documents and databases so the agent bases its answers on verified content and names the source – traceable instead of guessed.
A secure environment on GEAP
We set up your environment on the Gemini Enterprise Agent Platform: access rights, network, encryption and logging following Google's recommendations.
Operations and monitoring
We monitor your agents' answer quality, errors and costs, test new versions before rollout and keep improving them.
Clean data pipelines
So your agent works with current and correct data, we build the data pipelines behind it – with data protection and access rules from the start.
From practice: AI agents in sales
How a machine builder with around 100 people in sales reviews tenders faster and partly automates quote preparation – with Gemini Enterprise for the whole team and a dedicated quoting agent on GEAP.
Find out in 30 minutes where AI agents can make the biggest difference for you
- A first assessment of your use cases
- Gemini Enterprise, GEAP or both?
- Next steps
Frequently asked questions
Gemini Enterprise is Google's AI platform for companies: an app in which employees work with Gemini, search company knowledge from connected systems and use AI agents or build their own without code. Since April 2026, the former Google Agentspace is also part of it.
GEAP is Google's platform for building, running and monitoring your own AI agents and applications. Google introduced it in April 2026; services such as Agent Studio, Agent Search, Agent Runtime and Model Garden have run under its umbrella since then.
Gemini Enterprise is aimed at all employees and is licensed per user. GEAP is aimed at developer and data teams who build their own agents, and is billed by usage. Agents from GEAP can be made available in Gemini Enterprise.
On April 22, 2026, Google moved Vertex AI into the Gemini Enterprise Agent Platform (GEAP). The features were carried over and renamed – for example, Vertex AI Search is now Agent Search and Vertex AI Agent Engine is now Agent Runtime. If you look for Vertex AI today, you'll find the services in the Google Cloud console under Agent Platform. Google: Gemini Enterprise Agent Platform release notes
More than 200 models in Model Garden: Gemini, Anthropic's Claude models such as Claude Opus 5.5, and open models from other providers. OpenAI models have been available in preview since October 2026 (according to OpenAI).
Yes. Gemini Enterprise can be connected to Microsoft 365, for example to SharePoint and Outlook, and uses the existing permissions. You don't need to switch to Google Workspace for this.
Yes. GEAP is vendor-neutral and integrates into existing M365 and Azure data landscapes — alongside Copilot, without a platform switch. Especially in Microsoft-centric companies, a scoped pilot on the existing data is the typical starting point.
A near-production proof of concept on GEAP is typically ready within a few weeks — connected to your real data, with guardrails and defined success criteria. You then decide, based on measurable results, whether and how to scale. For qualified projects, we check whether Google can fund the PoC.
Gemini Enterprise is licensed per user per month, depending on the edition. GEAP is billed by actual usage, for example by processed tokens. You can also get the licenses through us – in most cases below list price.
Yes, but not via a ready-made standard connector. SAP data is added via SAP Business Data Cloud Connect for BigQuery or a custom connector; for processes such as quotes or orders, a dedicated agent is built on the Gemini Enterprise Agent Platform.