News · AI agents

Gemini Enterprise: pricing, editions and the agents it builds

Google launched Gemini Enterprise on 9 October 2025 as its AI agent platform for businesses, the successor to Google Agentspace, and on 8 October 2026 it announced the next step: a universal agent for work.

For your business, Google Gemini Enterprise means agents that work with your email, your CRM and your documents without any coding. Here is which agents it supports, what each edition costs, how it differs from Gemini in Google Workspace and when you need an agent of your own.

By 13 min read

The news

What is Gemini Enterprise and what has changed?

Short answer

Gemini Enterprise is Google’s AI platform for businesses: an agent-powered chat that searches and works across your company’s data (Google Workspace, Microsoft 365, Salesforce and more than 100 applications) while respecting each employee’s permissions. It includes ready-made Google agents, a no-code agent builder and central governance.

Gemini Enterprise was born on 9 October 2025. Thomas Kurian, CEO of Google Cloud, presented it as “the new front door for AI in the workplace”: a single place where every employee chats with the company’s documents, data and applications, and where agents are built, used and controlled (Google Cloud, 9 October 2025). It took the place of Google Agentspace, the enterprise search tool with agents that Google had unveiled in December 2024 and which is now part of Gemini Enterprise.

Whether you search for Gemini for enterprise, Gemini Enterprise AI or simply Google Gemini with a business licence, it is the same product. But do not confuse it with the consumer Gemini app on your phone or with Gemini inside Gmail and Docs: they are three different things, and I compare them below.

The platform has changed a great deal in 2026, as the timeline shows. The latest step came on 8 October at Gemini at Work 2026: the “Gemini agent”, which plans the work, connects to your systems and hands back the finished task (Google Cloud, 8 October 2026).

  1. 13 Dec 2024

    Google unveils Agentspace

    Enterprise search with Gemini agents. It is now part of Gemini Enterprise.

  2. 15 Jan 2025

    Gemini joins the Google Workspace plans

    Business and Enterprise plans include Google’s AI with no paid add-on.

  3. 9 Oct 2025

    Gemini Enterprise launches

    Google agents, a no-code builder, connectors and central governance. Business edition at $21 a month.

  4. 22 Apr 2026

    Vertex AI becomes Gemini Enterprise Agent Platform

    Announcements include an improved Agent Designer, long-running agents, an agent inbox and the Agent Marketplace.

  5. 1 Aug 2026

    Pay-as-you-go edition

    Generally available: you pay for what you use, from a single licence.

  6. 3 Sep 2026

    Workflow Builder, formerly Agent Designer, becomes generally available

    Scheduled or on-demand flows, connected to Workspace and third-party tools.

  7. 8 Oct 2026

    Gemini at Work 2026: the “Gemini agent”

    A universal agent and “coworker agents” with their own Workspace account.

According to the figures Google publishes, adoption among large companies is already high, including in Spain:

$21

starting price per seat per month for the Business edition

Google Cloud, checked 12 Oct 2026

+100

integrations with business applications in its catalogue

Google Cloud, checked 12 Oct 2026

≈ 90%

of Fortune 100 companies use Gemini Enterprise

Google Cloud, 8 Oct 2026

+1,000

custom Gemini Enterprise agents built at Orange Spain

Google Cloud, 8 Oct 2026

My take

Google is no longer selling a chatbot for employees: it is selling the place where your company’s agents live, its own, yours and third parties’. That turns the choice into an architecture decision, not a licensing one: whoever controls that place controls which data your agents see and how much they cost you.

What’s included

Which agents can your business build with Gemini Enterprise?

With Gemini Enterprise, your business can use three kinds of agent: the ones Google provides, the ones any employee builds without coding and the ones your technical team builds with code. All of them are published in the same app and controlled from the same console.

1. Ready-made Google agents

  • Deep Research: runs hundreds of searches across the web and the internal data you have access to, and delivers a report with sources.
  • Gemini Notebook for enterprise (NotebookLM): summarises and extracts information from long, complex documents.
  • Data Insights: answers questions about BigQuery data without SQL. Its page labels it as a preview.
  • Partner agents: since 22 April 2026, the Agent Marketplace lets you request agents from companies such as Accenture, Oracle or ServiceNow, which your administrator approves (Google Cloud).

2. No-code agents with Workflow Builder (formerly Agent Designer)

Workflow Builder is the agent builder for people who do not code. You describe what you want in plain language or assemble it on a visual canvas, review and approve each step before it runs, and add checkpoints where a person has to sign off. Since 3 September 2026 it has been available to all customers, with scheduled or on-demand flows (release notes). A typical example: every Monday, read the open Jira issues and the week’s customer emails, and post a summary sorted by priority in the team chat.

3. Code-built agents: Gemini Enterprise Agent Platform

The Gemini Enterprise Agent Platform is the evolution of Vertex AI and the technical side of the offer. It includes Agent Studio, a visual environment whose agents can be exported to code; the Agent Development Kit (ADK) for programming them; templates; and a catalogue of more than 200 models, including Gemini 3.1 Pro and Anthropic’s Claude (Google Cloud, 22 April 2026). It is paid for separately, by usage, and the agents you build there reach employees through the Gemini Enterprise app. This is where the most autonomous agents are built, what is known as agentic AI: Agent Platform supports agents that keep working for days.

Gemini Enterprise connectors: which systems the agents can reach

An agent is only as good as the data it can reach. The catalogue of Gemini Enterprise connectors includes, among others:

  • Email and calendar: Gmail, Google Calendar and Microsoft Outlook.
  • Documents: Google Drive, SharePoint, OneDrive, Box, Confluence and Notion.
  • CRM and customer service: Salesforce and ServiceNow in Standard and Plus; HubSpot and Zendesk in Business as well.
  • ERP: Oracle NetSuite and Microsoft Dynamics 365, in Standard and Plus.
  • Teamwork: Slack, Microsoft Teams and Jira.
  • Anything missing: a custom connector or your own MCP server, the standard I explain in my article on the Model Context Protocol.

What about SAP?

Google cited SAP as a connection example at the October 2025 launch, but the public connector catalogue, checked on 12 October 2026, does not include an SAP connector. If your ERP is SAP or industry-specific software, ask to see the connection working with your own data before you sign.

The latest: the Gemini agent and “coworker agents”

The Gemini agent works inside Workspace, Microsoft 365 and Slack. Coworker agents get their own Workspace account, with email, calendar and Drive, and only see what you share with them. Every action is recorded in an audit log under the agent’s name, and you can set a spending cap per project. Mind the timelines: the product page notes that features such as background tasks are in early access.

Pricing

Gemini Enterprise pricing: how much does each edition cost?

Google Gemini Enterprise pricing is per seat, per month. The Business edition starts at $21 (USD) per seat per month (Google Cloud); Gemini Enterprise Standard and Gemini Enterprise Plus launched from $30 per seat per month on an annual commitment (TechCrunch, 9 October 2025), and Google now refers buyers to its sales team for a quote. Prices checked on 12 October 2026; Google publishes them in US dollars.

Gemini Enterprise editions (checked on 12 October 2026)
CriterionBusinessStandardPlusPay-as-you-goFrontline
Price (USD)From $21 per seat per monthFrom $30 per seat per month in 2025 (annual); now via salesFrom $30 per seat per month in 2025 (annual); now via salesBy usage, minimum 1 licenceVia sales, as an add-on
UsersTeams of up to 300 (technical cap: 500)1 or more, no cap1 or more, no capFrom 1150 or more Standard or Plus seats
Storage per user25 GiB30 GiB75 GiBBy usage2 GiB
ConnectorsA selectionFull catalogueFull catalogueFull catalogueFull catalogue
Build no-code agentsYesYesYesYesUse only
Data Insights and your own code-built agentsNoYesYesYesOnly those IT provides
Enterprise compliance (data residency, customer-managed encryption keys)NoYesYesYesYes

Features and limits come from the editions documentation, updated on 9 October 2026. The pay-as-you-go edition has been available since 1 August 2026 and requires a billing account with monthly invoicing; it lets you set a spending limit. Business, Standard and Plus can be trialled free for 30 days. And the developer platform, Agent Platform, is not included: it is paid for separately by usage, according to Google’s FAQs.

Watch out for the name Gemini Business: until January 2025 it was a paid Google Workspace add-on; today Google uses Business for the small edition, Gemini Enterprise Business. If you are comparing with OpenAI, my article on ChatGPT Business covers what its equivalent plan includes.

Change on 17 September 2026

New or renewed online subscriptions to Standard and Plus no longer include Gemini Code Assist, the coding assistant. Existing subscriptions keep it until the end of their term. If your technical team relied on it, check before you renew.

Comparison

Gemini in Google Workspace or Gemini Enterprise? How they differ

Gemini in Google Workspace is the assistant built into Gmail, Docs, Meet and Drive, and it works mainly with Google data; Gemini Enterprise is a separate platform that connects to applications outside Google, hosts Google, in-house and third-party agents, and controls them from a single console.

Since 15 January 2025, Workspace Business and Enterprise plans have included Google’s AI with no add-on (Google Workspace). According to the plans page, Gemini is in Gmail on every plan and in Docs, Sheets, Drive and Meet from Business Standard upwards. And since 4 December 2025 there has been Workspace Studio, which builds no-code agents for Workspace tasks, with steps into apps such as Asana, Jira, Mailchimp or Salesforce.

Gemini in Workspace, Gemini Enterprise and a custom AI agent, criterion by criterion
CriterionGemini in WorkspaceGemini EnterpriseCustom AI agent
What it isThe assistant inside Gmail, Docs, Meet and Drive, plus the Gemini appA platform to search, build and govern agents across the whole companyAn agent designed for one specific process in your business
Which data it reachesMainly Google Workspace dataWorkspace, Microsoft 365 and more than 100 applications, respecting permissionsThe systems you choose, including those with no standard connector
AgentsWorkspace Studio: no-code agents for Workspace tasksGoogle, no-code, code-built and third-party agents, in one consoleOne or several, with your rules, limits and approvals
How you payIncluded in Workspace business plansPer seat per month, or by usageA design project plus model usage costs
Where it shinesPersonal productivity inside GoogleWhat is common to the whole company, with governanceWhat sets you apart from your competitors
LimitationLittle reach outside GoogleWhatever is not in the catalogue or does not fit a templateNeeds design, testing and maintenance

The question is not which one is better, but where your work lives. If your company runs on Gmail and Drive, start with what you already pay for. If your data is spread across Microsoft 365, a CRM and an ERP, the platform makes sense; and if you live in Microsoft, compare it with Copilot Studio for building agents in Microsoft 365.

Gemini Enterprise is very good at what every company has in common. What sets you apart from your competitors is hardly ever common.

My take · José Galán

Data and compliance

Is Gemini Enterprise available in the EU, and does it keep data in Europe?

Yes: the Standard and Plus editions can be set up in the eu multi-region, which stores data at rest inside the European Union and processes the models in Europe, with exceptions worth knowing (locations documentation, updated on 9 October 2026).

Google recommends the global location unless you have regulatory reasons, because new models and features arrive there first. If you choose eu, the following is left out:

  • Image generation with Nano Banana Pro, video generation and answers grounded in Google Search only work in the global region.
  • Agents’ code execution does not guarantee that data at rest stays in the EU.
  • In Gemini Notebook for enterprise, the content studio and the search for new sources do not meet in-Europe processing.
  • Customer-managed encryption keys (CMEK) only work in the us or eu multi-regions, not in global.

What about Gemini Pro?

This is the most important caveat for a European business. The latest version of Gemini Pro covered in the documentation, Gemini 3.1 Pro, is in limited availability and meets neither residency at rest nor processing in the EU. Gemini 2.5 Pro and the recent Flash models do. In other words: if you require data in Europe, you may not be working with Google’s newest models.

Two more points. Google states that your data, prompts and outputs are not used to train its models or other customers’ models (product page). And the Business edition does not include enterprise-grade compliance: it is not the option if you need data residency.

The region does not make you compliant

Choosing eu helps with the GDPR, but it does not solve everything: you still need a lawful basis, a record of processing activities and a clear view of what each agent does. And if an agent makes decisions about people, the EU AI Act adds obligations. I explain this in my guide to the EU AI Act and the GDPR for SMEs.

My recommendation

When do you need a custom AI agent instead of Gemini Enterprise?

You need a custom AI agent when the value lies in a decision specific to your business, in a system the catalogue does not cover or in a customer-facing channel; for everything that is common, Gemini Enterprise, or even Gemini in Workspace, is usually enough.

My recommendation, as José Galán, Search & AI Visibility consultant and AI agent designer, is to draw up two lists before you buy anything. Gemini Enterprise fits when:

  • The problem is finding information spread across many applications.
  • You want each team to build simple agents under IT control.
  • The tasks look the same in any company: reports, summaries, drafts, CRM queries.

A custom agent, on the other hand, is usually the better option when:

  • The task needs your business judgement: which payment to chase first, which product line is about to run out, which customer is at risk.
  • The data lives in an ERP or industry-specific software with no standard connector.
  • The agent serves customers. Gemini Enterprise is built for employees; for customer service, Google offers a different product, its Customer Engagement Suite.
  • You need rules, limits and approvals that do not fit a template.

If you are still torn between an agent, a chatbot or an automation, my guide AI agent vs chatbot vs automation helps you decide for your own case. And if you are already clear, I design custom AI agents that analyse your data, prioritise and explain why, with a person approving what matters. Tick what you need here and see what fits:

Gemini in Workspace, Gemini Enterprise or a custom AI agent?

1 · What does your company work with?
2 · Tick what you really need
Gemini in Workspace—
Gemini Enterprise—
Custom AI agent—

Tick at least one need to see the recommendation.

A rough guide: it scores your answers against the criteria in this article. Nothing is stored or sent.

You do not have to pick just one

In Standard and Plus, Gemini Enterprise lets you use your own code-built agents and register external agents through the A2A protocol. A custom agent can live inside the same app your staff use, with the same permission controls.

Impact

What does Gemini Enterprise mean for your business?

It means Google’s AI stops being just a writing assistant and becomes an agent platform that touches your systems. The impact varies by area:

LeadershipHigh impact

Decide the architecture before the tool: what the shared platform will do and which processes deserve an agent of their own. If you want an outside view, that is what AI consulting for businesses is for.

MarketingMedium impact

Deep Research and no-code agents speed up reports and briefs. Getting Google’s AI, including Google AI Mode, to cite your brand to the outside world is a different job: AI SEO.

Sales and customer serviceMedium impact

With Salesforce or HubSpot connected, the team can query the CRM in plain language. But an agent that serves customers on your website or on WhatsApp takes a different route.

IT and dataHigh impact

Governance falls to you: each agent’s identity, least-privilege permissions, the eu region, spending limits and which connectors are switched on. Review how your own agents are published, too.

Action plan

What should you do now?

Start with what you already pay for and with measurable processes; choose the platform afterwards. These are the steps for this week and this quarter:

  1. This week: take stock of what you already pay for

    If you have Google Workspace, you already have Gemini and Workspace Studio. If your team uses ChatGPT or Microsoft 365 Copilot, note that down too. Do not pay twice for the same thing.

  2. This week: pick two or three processes and map them

    Write down how they are done today, how long they take and which data they touch. It is the first step of the MAPA method for mapping and automating processes with AI; you will find the MAPA template and the “Can AI read your website?” checklist among my free resources.

  3. This month: run a 30-day trial with a pilot team

    Business, Standard and Plus have a 30-day free trial. Measure the same things before and after in those processes, not general impressions.

  4. This quarter: settle data, region and compliance

    Choose the edition by your requirements: if you need data in the EU, Business is not enough. Set up the eu region from the start, because data stores are created in a specific location.

  5. This quarter: decide where you need your own agent

    For the process that sets you apart, design an agent with your rules and a person who approves what matters. If you use Gemini Enterprise, it can live inside it.

Key points

  • Gemini Enterprise has been Google’s agent platform for businesses since 9 October 2025 and took the place of Google Agentspace.
  • It brings Google agents, a no-code builder (Workflow Builder) and connections to more than 100 applications, including Microsoft 365.
  • The Business edition starts at $21 per seat per month; Standard and Plus are quoted by sales and can be trialled for 30 days.
  • For data in the EU you need an edition other than Business, and Gemini 3.1 Pro does not yet meet that residency.
  • The platform handles what is common; what sets you apart usually calls for a custom AI agent.

FAQ

Frequently asked questions about Gemini Enterprise

What is Gemini Enterprise?

Gemini Enterprise is Google’s AI agent platform for businesses, launched on 9 October 2025 as the successor to Google Agentspace. It brings together in one app a chat that searches company data while respecting each employee’s permissions, Google agents such as Deep Research, a no-code agent builder called Workflow Builder and a console to govern every agent, including your own and third-party ones.

What is the Gemini Enterprise pricing for each edition?

The Business edition starts at $21 (USD) per seat per month and comes with a 30-day free trial. Standard and Plus launched in October 2025 from $30 per seat per month on an annual commitment; Google now refers buyers to its sales team for a quote. There is also a pay-as-you-go edition from a single licence, and Gemini Enterprise Agent Platform is billed separately, by usage. Prices checked on 12 October 2026.

What is Gemini Enterprise Business, and is it the same as Gemini Business?

Today, Gemini Enterprise Business is the small edition of Gemini Enterprise: designed for teams of up to 300 people, with no IT setup, a selection of connectors, Deep Research and no-code agents. It does not include the full connector catalogue, Data Insights, your own code-built agents or enterprise-grade compliance such as data residency. Until January 2025, Gemini Business was also the name of a paid Google Workspace add-on.

Does Google Workspace with Gemini include agents, or do I need Gemini Enterprise?

Google Workspace business plans already include Gemini: in Gmail on every plan, and in Docs, Drive and Meet from Business Standard upwards. They also include Workspace Studio, which builds no-code agents for Workspace tasks with steps into apps such as Asana, Jira or Salesforce. You need Gemini Enterprise when you want to work with data from many other applications, use your own or third-party agents and govern them all from a single console.

What is the Gemini Enterprise Agent Platform?

Gemini Enterprise Agent Platform is the evolution of Vertex AI that Google announced on 22 April 2026. It is the technical layer: Agent Studio and the Agent Development Kit for building agents with code, more than 200 models including Gemini and Claude, and services to deploy, govern and evaluate them. It is paid for by usage on Google Cloud, and its agents reach employees through the Gemini Enterprise app.

Does Gemini for enterprise keep data in the EU?

It can. The Standard and Plus editions let you choose the eu multi-region, which stores data at rest inside the European Union and processes the models in Europe. There are exceptions: Gemini 3.1 Pro, video generation and answers grounded in Google Search do not meet that residency. Google also states that your data is not used to train its models. The Business edition does not include these controls.

When is a custom AI agent better than Gemini Enterprise?

When the value lies in something specific to your business: a decision made with your own rules, such as prioritising collections or forecasting stock-outs; a system that is not in the connector catalogue, such as an industry-specific ERP; or a customer-facing channel such as WhatsApp, because Gemini Enterprise is built for employees. It is not either-or: in Standard and Plus you can publish your own agent inside Gemini Enterprise and govern it from there.

References

Sources

Checked and verified on 12 October 2026.

This article was created with the help of AI and reviewed by José Galán. I take great care over every post and every translation, but the odd mistake can still slip through. If you find one, write to me: you will be helping me improve.

Custom AI agents

Does Gemini Enterprise fall short for your process? Let’s look at it together.

I design internal AI agents that analyse your data, prioritise and spot risks, with an assistant that explains why. Data in the EU, and you approve every important action. If Gemini Enterprise or Workspace is enough for you, I will tell you that too.

See how I work

Written by

Agent-as-a-Service Designer at AllHub and Search & AI Visibility consultant. I design internal AI agents and help companies show up in Google and AI engines, with reproducible measurement.

LinkedInBackground

Shall we talk?

Get recommended by AI and/or put an agent to work for you.

Tell me what you want to achieve: fill in the form or email me at info@josegalan.dev.

We go through your AI visibility snapshot (or the process you want to automate) and I tell you, with data, what I would change first. No commitment.

What are you interested in?
Company size (optional)
When would you like to start? (optional)

Basic data protection information. Controller: José Galán. Purpose: to reply to your enquiry. Legal basis: your consent, which you can withdraw at any time. Recipients: no data is shared with third parties unless required by law; Google (hosting and form delivery) acts as data processor. Your rights: access, rectification, erasure, objection, restriction and portability. More information in the privacy policy.