Prepare and test your skills
Prepare and test your skills
Custom Gems, role-based assistants, and the Gemini Enterprise app are prebuilt ways to bring generative AI into daily work without building models. Gems provide reusable expertise for repeated tasks, Gemini for Google Cloud provides assistants inside developer and data tools, and the Gemini Enterprise app provides chat, drafting, and agents for knowledge workers. Together they trade custom development for speed and consistency.
Custom Gems are reusable AI experts inside Gemini Advanced that save a role, tone, and set of rules for a recurring job. A Gem holds the background instructions once, such as how to review contracts, coach support replies, or summarize research, and applies that same pattern every time it is used. This reduces setup time and keeps answers aligned to the same standards when different people need the same specialized behavior.
Gemini in Looker (opens in a new tab) is part of a suite that divides help by where the user already works. A data analyst gets help in BigQuery or Looker, a developer gets help in Code Assist, and a security analyst gets help in Security Command Center. The main offerings differ in audience and daily tasks:
| Product | Type of assistance | Example tasks |
|---|---|---|
| Gemini Cloud Assist | Design, optimize, and operate cloud applications | Manage applications, troubleshoot resources, manage security configurations |
| Gemini Code Assist | Develop, deploy, and troubleshoot code | Generate and debug code, generate unit tests, build integrations |
| Gemini in BigQuery | Prepare and analyze data | Data canvas, data insights, SQL and Python generation |
| Gemini in Looker | Analyze and export Looker data | Custom visualizations, LookML, conversational analytics |
| Gemini in Security Command Center | Simplify search and interpret cases | Generate search queries, summarize security cases |
Gemini Code Assist also builds enterprise integrations from natural language prompts, including connector tasks, Call REST API tasks, JavaScript tasks, and control flow such as For Each Loop and While Loop. It cannot generate sub-integrations.
The Gemini Enterprise app (opens in a new tab) is a web and mobile assistant for knowledge workers that answers questions from connected company data. Data flows from company sources into the app, and the app returns verified answers, drafts, and links to agents. Users can draft and format documents in Canvas while continuing the chat, generate images for documents, and build private notebooks grounded in uploaded files, Google Drive files, and web links.
The mobile app (opens in a new tab) extends the same work with camera photos as context, file uploads, voice input, audio answers, and interactive cards for actions such as sending email, which the user reviews and confirms. Users connect to agents through the Agent Gallery, which holds Google-made agents, organization-shared tools, and custom multi-step agents, and they can install custom skills that standardize recurring workflows.
Gemini Notebook Enterprise and Gemini Enterprise are two different answers to knowledge work. Gemini Notebook Enterprise is a focused research assistant that answers only from a curated notebook of documents, websites, and notes. Gemini Enterprise is a broad search and agentic platform that searches Google and third-party systems and uses agents to act across systems.
| Dimension | Gemini Notebook Enterprise | Gemini Enterprise |
|---|---|---|
| Primary goal | Centralized knowledge from your own documents | Enterprise-wide search and agent orchestration |
| Input | Uploaded documents and curated sources | Broad queries across enterprise systems |
| Output | FAQs, timelines, study guides, audio podcasts, and Q&A confined to sources | Answers, images, videos, reports, and automated actions |
| Choose when | Starting sources are known and authoritative | Starting point is unknown or work needs agents |
Combined use follows a fixed order: first, Gemini Enterprise searches broadly and surfaces valuable sources; then, the user adds those sources into a notebook; finally, future chat and content stay grounded in that curated set.
Gemini in BigQuery (opens in a new tab) shows how prebuilt AI must be balanced with safety controls for live use. Model Armor screens Gemini Enterprise requests and responses before they reach their destination. Traffic flows from Gemini Enterprise to Model Armor for checking, and Model Armor replies according to its template, so Gemini Enterprise either blocks or allows the request to proceed. For example, if a request contains personally identifiable information (PII) and the template blocks PII, Model Armor tells Gemini Enterprise to block it.
Secure posture controls (opens in a new tab) add two more layers. Organization policies control settings at creation time on new Vertex AI Workbench resources, such as turning off file downloads, root access, terminal access, and public IP access, and requiring an automatic upgrade schedule. Custom Security Health Analytics detectors report risks on resources that already exist, such as a dataset or model not encrypted with a customer-managed encryption key (CMEK). The order is fixed: Model Armor filters live traffic, policies control creation, and detectors report on deployed resources.
Exam tip: Organization policies prevent insecure settings on newly created Vertex AI Workbench resources, while Security Health Analytics detectors report CMEK risks on datasets and models that already exist.
Gemini Enterprise (opens in a new tab) is an intranet search, AI assistant, and agentic platform that connects Google and third-party systems into one permissions-aware interface. Employees ask once in text or voice, and the platform searches connected sources and returns grounded answers, summaries, images, or video. The business value is faster discovery, fewer tool switches, and decisions built from complete information.
Enterprise-wide multimodal search (opens in a new tab) is a search layer that sits over connected company systems. Data flows from Google and third-party sources through prebuilt connectors into dedicated data stores, and a Gemini Enterprise app serves results from those data stores to the user. Supported connections include Confluence, Jira Cloud, Microsoft SharePoint, ServiceNow, plus Google content and other SaaS systems.
User query flows to the Gemini Enterprise app, which checks SSO and user-level access at a trust boundary, ingests Google, Confluence, Jira, SharePoint and ServiceNow content via connectors into data stores, and returns grounded answers with citations.
The trust boundary sits between the user and the stored data. Traffic flows from the user query to the app, the app checks single sign-on (SSO) and user-level access, and only content the user is allowed to see flows back as results. Because answers are grounded in those data stores, they include citations to enterprise sources, with optional grounding in public website data when enabled.
Cross-media retrieval and generation is the ability to read one media type and answer or create in another. The in-app assistant analyzes uploaded PDFs and images, summarizes complex files, answers follow-up questions, generates images from text prompts, and creates reports with streamed research. A user can upload a specific document for a one-off question or ask the assistant to analyze Excel files, CSV files, and images inside PDFs.
Common work tasks show the range:
Choose this capability when the task needs synthesis from many sources rather than lookup in one known file.
Enterprise grounding (opens in a new tab) means answers are built from connected organizational data rather than from the model alone. The Gemini Enterprise assistant generates answers grounded in data store content with citations, which improves accuracy and keeps knowledge management focused on approved sources. A customizable retrieval-augmented generation (RAG) agent can stay grounded on enterprise data across many formats, geographies, and languages.
Gemini Notebook Enterprise provides the focused side of the same strategy. A notebook is an organized collection of curated documents, websites, and notes that becomes the single trusted reference for a project or team. Chat, FAQs, timelines, study guides, and audio podcasts stay strictly confined to that notebook, and data remains under access control and is not used for external model training.
Knowledge flows from broad search into focused curation in order:
Exam tip: Choose Gemini Notebook Enterprise when answers must stay confined to a known set of sources, and choose Gemini Enterprise when the starting point is unknown and search must span all enterprise data.
Custom agents are applications that accomplish specific objectives by connecting to enterprise data and using tools to automate multi-step processes. Within Gemini Enterprise, an app provides search results, actions, and agents to users, and the app must connect to one or more data stores to use their data. Actions let the platform move beyond retrieval to create or update records in real time, such as creating Google Calendar events or editing Jira Cloud issues. Skills provide reusable instructions that customize task completion without code.
Teams can build agents with Workflow Builder, start from prebuilt agents in the Agent Gallery, or import outside agents through Agent Registry or an A2A card. When importing, the administrator follows an ordered setup: open the target app and choose Agents > Add agent, choose Agents from Agent Registry or Custom agent via A2A, provide details such as name, agent URL, and skills, authenticate with Default Google-managed credentials or Custom OAuth, finish setup so the agent shows state Enabled and type A2A (Custom), and share it with users or groups.
Users invoke enabled agents by browsing the Gallery, opening a direct link, using @mention in core chat, or letting Gemini Enterprise route a question to the relevant data agent. After a one-time OAuth sign-in to the data source, the agent streams back text, Markdown, charts, or tables, and the conversation is saved in history. Data agents for BigQuery (opens in a new tab) support this natural-language analysis over governed data.
Observability and analytics keep agents governed and observable as they run. Observability provides real-time telemetry with end-to-end request traces, trace spans, and usage audit logs to diagnose latency and monitor step execution. Analytics provides usage trends, search quality, and engagement through an interactive dashboard powered by Looker.
Teams can preview an agent, toggle a debug panel to see reasoning, tool calls, and retrieval steps, and use the Tools menu to connect external APIs or private data stores. Gemini Enterprise for CX (opens in a new tab) shows the full lifecycle in practice, from creation to experimentation, evaluation, deployment, management, human supervision, and self-optimization, with Agent Studio for self-service, Agent Assist for live coaching, Insights for operations data, and Commerce agents that connect chat and voice to backend tools.
Gemini for Google Workspace (opens in a new tab) is a generative AI-powered collaboration product that assists business users inside everyday work apps. Instead of switching to a separate tool, the user describes what is wanted in natural language, reviews the draft Gemini creates, and edits the final version. The result is faster creation, easier creativity, and smoother teamwork.
Gemini for Workspace generates original content such as text, summaries, images, and organized data from a natural language request. Like Gemini in Cloud products, which helps developers and data scientists inside their consoles, Gemini for Workspace provides one continuous experience inside Gmail, Docs, Slides, Sheets, and Meet. The request flows from the open Workspace app to the Gemini models, and the response flows back into the same document, message, spreadsheet, presentation, or meeting.
Core assistance applies the same pattern across five apps: describe the need, review the starting point, then edit and approve. Drafting fits Gmail and Docs, visual creation fits Slides, organizing data fits Sheets, and real-time summaries fit Meet:
| App | Type of assistance | Example tasks |
|---|---|---|
| Gmail | Drafting and summarizing communication | Draft replies, summarize threads, adjust tone and length |
| Google Docs | Drafting and refining content | Draft from a prompt, summarize, rewrite and proofread |
| Google Slides | Creating visual content | Generate images, draft outlines, summarize slides |
| Google Sheets | Organizing and analyzing data | Organize tables, generate formulas, summarize patterns |
| Google Meet | Real-time meeting assistance | Capture summaries, highlight key points, support follow-up |
Exam tip: Choose Gmail or Docs for drafting and summarizing text, Slides for images and presentation content, and Sheets for organizing data into tables and formulas.
Context-aware processing (opens in a new tab) combines a user request with relevant work context to return a ready-to-review answer. The input a user submits is called a prompt, and the answer Gemini returns is called a response. Gemini in the Cloud console can include basic context such as project ID and organization ID, while Gemini in BigQuery can access table metadata, schema, sampled data, and job histories. The response moves in order:
The same pattern supports practical work. Automated summarization condenses cases, code functions, or tables into short summaries with next steps. BigQuery data canvas (opens in a new tab) lets users find, join, query, and visualize tables with natural language in one collaborative flow. Conversational analytics uses data agents with custom metadata, glossary terms, and verified queries so answers stay consistent with business definitions. Gemini in Google Security Operations (opens in a new tab) generates search queries, summarizes cases, suggests resolution steps, and answers threat-intelligence questions.
Enterprise-grade privacy for Gemini rests on encryption, identity controls, and limits on training. Gemini does not use prompts or responses to train models, and data sent to models is used strictly to serve the response. Prompts are encrypted in transit, and data is encrypted at rest and in transit.
Protection uses Google Cloud controls:
Planning must account for limits. Gemini output can seem plausible but be factually incorrect, so users should validate all output. Setup choices affect compliance (opens in a new tab), because Gemini in BigQuery does not support the same compliance offerings as BigQuery and does not provide data residency for individual locations. Gemini processing can be specified for US- and EU-supported jurisdictions, while data outside those jurisdictions is processed globally.
A custom Gem is a reusable AI expert inside Gemini Advanced that saves a role, tone, and set of rules for a recurring job. It holds the background instructions once, such as how to review contracts, coach support replies, or summarize research, and applies that same pattern every time it is used to reduce setup time and keep answers aligned to the same standards.
Gemini Notebook Enterprise is a focused research assistant that answers only from a curated notebook of documents, websites, and notes, while Gemini Enterprise is a broad search and agentic platform that searches Google and third-party systems and uses agents to act across systems. Choose Gemini Notebook Enterprise when starting sources are known and authoritative and answers must stay confined to those sources, and choose Gemini Enterprise when the starting point is unknown or work needs agents.
Enterprise-wide multimodal search sits over connected company systems where data flows from Google and third-party sources through prebuilt connectors into dedicated data stores, and a Gemini Enterprise app serves results from those data stores to the user. The app checks user-level access so only content the user is allowed to see flows back as results, and answers are grounded in those data stores with citations to enterprise sources.
Gemini for Google Workspace is a generative AI-powered collaboration product that assists business users inside everyday work apps by generating original content from a natural language request. The request flows from the open Workspace app to the Gemini models and the response flows back into the same document, message, spreadsheet, presentation, or meeting, such as drafting replies in Gmail and Docs, generating images and outlines in Slides, organizing tables and formulas in Sheets, and capturing summaries in Meet.
Last updated