Generative AI Leader · Product map
Google AI product map 2026: old and new names
· approx. 11 min read
Independent study aid – not by Google, not affiliated with Google.
Since April 2026, Vertex AI has officially been called Gemini Enterprise Agent Platform (Agent Platform for short). Agentspace became Gemini Enterprise, and NotebookLM became Gemini Notebook in July 2026. This map sorts all Google Cloud gen AI offerings into the five layers infrastructure, models, platforms, agents and applications – with former names, audience and typical use.
Renames at a glance
Older courses and blog posts often still use the old names. In the exam, the name from the Exam Guide counts; the current documentation sometimes already uses newer names.
| Exam Guide (authoritative) | Study Guide / current docs | formerly | Evidence |
|---|---|---|---|
| Gemini Enterprise Agent Platform / "Agent Platform" | Agent Platform | Vertex AI (until April 2026), before that AI Platform | release notes (2026-04-22) |
| Agent Search (on Gemini Enterprise Agent Platform) | Agent Search | Vertex AI Search (before that Enterprise Search / Gen App Builder) | release notes |
| Agent Studio | Agent Studio | Vertex AI Studio | release notes |
| Agent Platform AutoML | Agent Platform AutoML | Vertex AI AutoML | release notes |
| Agent Platform Feature Store | same | Vertex AI Feature Store | Study Guide p. 9 |
| Model Garden | Model Garden | Vertex AI Model Garden | release notes |
| RAG APIs | RAG Engine, Vector Search | Vertex AI RAG Engine, Vertex AI Vector Search | release notes |
| Gemini Notebook (API) | Gemini Notebook, Gemini Notebook Enterprise | NotebookLM (renamed 2026-07-16) | blog.google |
| Gemini Enterprise | Gemini Enterprise | Google Agentspace (until Oct. 2025) | Study Guide p. 5; own classification for the old name |
| Customer Engagement Suite | Gemini Enterprise for Customer Experience (CX) | Contact Center AI (CCAI) | Study Guide p. 5 |
| Conversational Agents | Customer Experience (CX) agents / CX Agent Studio | Dialogflow CX | Study Guide p. 5 |
| Conversational Insights | Customer Experience Insights (CX Insights) | CCAI Insights | Study Guide p. 5 |
| Agent Assist | Agent Assist | – | Study Guide p. 5 |
| Google Cloud Contact Center as a Service | CCaaS | CCAI Platform | Study Guide p. 5 |
| Gemini for Google Workspace | Google Workspace with Gemini | Duet AI for Workspace | Study Guide p. 5 |
| Gemini app + Gemini Advanced | Gemini app, Gemini Advanced | Bard (2023) | Study Guide p. 5; own classification for Bard |
| Cloud Functions (EG 2.5) | Cloud Run functions | Cloud Functions (renamed 2024) | cloud.google.com/functions |
| Imagen (EG §1.4) | Imagen; in Gemini products also the image model "Nano Banana" | – | Exam Guide §1.4; own classification |
Discrepancy: For the customer service bundle, the Exam Guide is older than the Study Guide. It names Customer Engagement Suite, Conversational Agents and Conversational Insights. The Study Guide names Gemini Enterprise for Customer Experience, Customer Experience Agent Studio and Customer Experience Insights. Both names can appear in the exam. Learn them as pairs.
Gemini Advanced has been called "Google AI Pro" in the consumer space since 2025. The Exam Guide still uses "Gemini Advanced". own classification
A. Applications: ready-made gen AI applications
| Product (current) | Formerly | Layer | Audience | Typical use case | Choose it when … |
|---|---|---|---|---|---|
| Gemini app | Bard | Applications | Individuals, business users | Writing, summarizing, translating, images, research | a single person needs quick, general AI help |
| Gemini Advanced | (consumer offering today: "Google AI Pro") | Applications | Power users, enterprises | Coding, complex reasoning, latest models | top model quality, priority access to new features and enterprise protection are required |
| Gems (in Gemini) | – | Applications | Business users | Custom assistant with a role, instructions, files | a task recurs and needs a consistent style or isolated chats |
| Saved Info (in Gemini) | – | Applications | Business users | Gemini deliberately remembers specific facts about the user | a certain context should apply permanently, without the entire history |
| Gemini for Google Workspace (SG: Google Workspace with Gemini) | Duet AI for Workspace | Applications | All employees | Gmail drafts, Docs, Slides images, Sheets formulas, Meet captions and summaries, Vids | AI should help directly inside office apps |
| Gemini Notebook (Pro / Enterprise) | NotebookLM (until 2026-07-16) | Applications | Knowledge workers, learners, teams | Research, onboarding, knowledge base, Audio Overview, only from your own sources | answers should be based exclusively on uploaded sources. Enterprise when IAM and compliance matter |
| Gemini Notebook API | NotebookLM Enterprise API | Applications/Platform | Developers, IT | Create and manage notebooks programmatically | notebook capabilities should be integrated into business processes |
| Gemini Enterprise | Google Agentspace | Applications (+ Agents) | Employees across the whole company | Search across all company data sources, assistants with connectors, custom agents, Notebook Enterprise | employees need a central, secure agent home page across all internal systems |
| Gemini for Google Cloud (Code Assist, Cloud Assist, in BigQuery, Looker, Databases, Colab Enterprise, Security) | Duet AI for Google Cloud | Applications | Developers, admins, analysts, security teams | Writing code, operating the cloud, analyzing data, investigating threats | technical teams need AI help inside Google Cloud tools |
| Google Search (Grounding with Google Search) | – | Applications/Models | End users; developers via grounding | Up-to-date web information | answers need current world knowledge beyond the knowledge cutoff |
| Google Vids | – | Applications | Business users | Create a first video draft | internal videos need to be produced quickly |
B. Customer experience: customer-facing suite
| Product (EG) | Current SG / docs name | Formerly | Layer | Audience | Typical use case | Choose it when … |
|---|---|---|---|---|---|---|
| Customer Engagement Suite | Gemini Enterprise for Customer Experience (CX) | Contact Center AI (CCAI) | Applications (+ Agents) | Contact centers, CX teams | End-to-end package for customer interactions | the company wants to modernize customer service with AI |
| Conversational Agents | CX agents / CX Agent Studio | Dialogflow CX | Agents | Contact centers, CX developers (low-code) | Chat and voice bots that talk to customers themselves; deterministic, generative, hybrid | customers should get 24/7 self-service |
| Agent Assist | Agent Assist | – | Applications | Human contact center agents | Live suggestions, knowledge, live transcription, translation, summarization | a human leads the conversation and needs real-time support |
| Conversational Insights | Customer Experience Insights (CX Insights) | CCAI Insights | Applications | CX management, QA | Analysis of all conversations, topics, sentiment, Generative FAQ | you want to understand trends and quality across all interactions |
| Google Cloud Contact Center as a Service | CCaaS | CCAI Platform | Applications/Infrastructure | Contact center operations | Omnichannel contact center, CRM integration, managed infrastructure | a complete cloud-native contact center is needed |
C. Platforms: building for developers and AI practitioners
| Product (current) | Formerly | Layer | Audience | Typical use case | Choose it when … |
|---|---|---|---|---|---|
| Gemini Enterprise Agent Platform ("Agent Platform") | Vertex AI (2021–04/2026), AI Platform, ML Engine | Platforms | Developers, AI practitioners | Build, tune, deploy and operate models and agents (fully managed) | you want to develop your own gen AI or ML solutions on Google Cloud |
| Model Garden | Vertex AI Model Garden | Platforms (catalog for models) | Developers, AI practitioners | Discover, test, deploy and tune over 200 models (according to the Agent Platform overview) | you need an existing model (Google, open, third-party) |
| Agent Studio | Vertex AI Studio | Platforms | Professional developers | Test and save prompts, tune models, low-code agent design, production-ready with cloud governance | enterprise prototyping through to production with IAM, security, quotas |
| Google AI Studio | (MakerSuite) | Platforms (developer tool outside the cloud) | Beginners, students, researchers, developers | Try Gemini for free, parameter playground, API key | fast, free prototyping with the Gemini Developer API |
| Agent Search (on Gemini Enterprise Agent Platform) | Vertex AI Search (Enterprise Search, Gen App Builder) | Platforms/Agents | Developers, digital teams | Search and recommendation experiences over your own data; prebuilt RAG; Document, Media, Healthcare, Commerce Search | customers or employees should get Google-quality search with grounding over company data, with little effort |
| RAG APIs (RAG Engine, Vector Search, Grounding APIs) | Vertex AI RAG Engine, Vertex AI Vector Search | Platforms | Developers | Custom RAG pipeline with control over retrieval | prebuilt RAG is not enough and full control is wanted |
| Agent Platform AutoML | Vertex AI AutoML | Platforms | Analysts, developers without deep ML expertise | Custom models for image, video, tabular data (classification, regression, forecasting) | you have your own data but no ML experts |
| Custom training (Model Builder) | Vertex AI custom training | Platforms | ML engineers, researchers | Full control over architecture and training | maximum control or novel models are needed |
| Agent Platform Feature Store | Vertex AI Feature Store | Platforms (MLOps) | AI practitioners | Share and reuse features, consistency between training ↔ serving | several models need the same prepared data |
| Agent Platform Model Registry | Vertex AI Model Registry | Platforms (MLOps) | AI practitioners | Versions, changes, lifecycle | versioning and traceability are required |
| Model Evaluation | Vertex AI Model Evaluation | Platforms (MLOps) | AI practitioners | Assess and compare models | you have to pick the best of several models |
| Agent Platform Pipelines | Vertex AI Pipelines | Platforms (MLOps) | AI practitioners | Workflow orchestration, from data preparation to deployment | the entire ML pipeline should be automated |
| Model Monitoring on Agent Platform | Vertex AI Model Monitoring | Platforms (MLOps) | AI practitioners | Detect performance degradation, skew and drift, trigger retraining | a production model gets worse |
| Agent Development Kit (ADK) | – | Platforms/Agents | Developers | Build agents in code, multi-agent systems | complex, custom-coded agents are needed |
| Agent Runtime | Vertex AI Agent Engine | Platforms | Developers | Run agents as a managed service (sessions, Memory Bank) | your own agents should run at scale in production |
| Apps Script | – | Platforms (low-code) | Power users, developers | Automate Workspace, Gemini add-ons (JavaScript) | Workspace processes should be automated with little code |
| AppSheet | – | Platforms (no-code) | Business departments, citizen developers | Business apps without code, with Gemini | business users want to build an app without code |
| Gemini API | PaLM API | Platforms | Developers | Integrate Gemini via HTTP into any app | your own application should call Gemini |
D. Agent tooling: cloud services and pre-built AI APIs (§2.5)
| Product | Formerly | Layer | Audience | Typical use case | Choose it when … |
|---|---|---|---|---|---|
| Cloud Storage | – | Infrastructure (tool) | Developers | Object storage for unstructured data, e.g. documents for search/RAG | an agent needs to read or store large files |
| Databases (Cloud SQL, Spanner, Firestore; BigQuery) | – | Infrastructure (tool) | Developers | Structured data, user data, agent progress | an agent needs to store or retrieve structured information |
| Cloud Run functions | Cloud Functions | Infrastructure (tool) | Developers | Small, event-driven serverless function as a tool | a single action should run in response to an event |
| Cloud Run | – | Infrastructure (tool) | Developers | Run containers serverlessly | a tool needs special dependencies or more control |
| Speech-to-Text API | – | Models (pre-built API) | Developers | Transcribe calls and meetings | speech → text |
| Text-to-Speech API | – | Models (pre-built API) | Developers | Voice UI, reading text aloud | text → speech |
| Translation API | – | Models (pre-built API) | Developers | Translate text, websites, audio and video | content needs to be translated |
| Document Translation API | – | Models (pre-built API) | Developers | Translate formatted documents | the layout must be preserved during translation |
| Document AI API | – | Models (pre-built API) | Developers | Extract data from invoices, forms, PDFs | structured data from documents is needed |
| Cloud Vision API | – | Models (pre-built API) | Developers | Image analysis, OCR, moderation, visual search | image content needs to be understood |
| Cloud Video Intelligence API | – | Models (pre-built API) | Developers | Analyze videos, make them searchable | video content should be analyzed |
| Natural Language API | – | Models (pre-built API) | Developers | Sentiment, entities, classification | text analysis (e.g. customer sentiment) is needed |
| Google Cloud API Library | – | Platforms | Developers | Further APIs (Maps, Workspace, YouTube …) | an agent should interact with other Google services |
E. Models: Google foundation models (§1.4, listed here for orientation)
| Model | Layer | Typical use case | Choose it when … |
|---|---|---|---|
| Gemini (Pro / Flash / Flash-Lite / Nano) | Models | Multimodal understanding, conversation, content, Q&A | general purpose, multimodal. Pro for complex tasks, Flash for fast and low-cost ones (Pro costs several times as much per token, see the pricing page). Nano on-device |
| Gemma | Models (open) | Open, customizable models, local deployments | you want to host, run locally or customize the model yourself |
| Imagen (alternative: Gemini image models "Nano Banana") | Models | Text-to-image (diffusion) | high-quality images should be generated |
| Veo | Models | Video from text or image | videos should be generated |
| Chirp | Models | Speech recognition (audio) | multilingual speech recognition |
| Lyria | Models | Generate music | music should be created |
F. Infrastructure and security
| Product | Formerly | Layer | Audience | Use case | Choose it when … |
|---|---|---|---|---|---|
| TPU | – | Infrastructure | AI practitioners | AI accelerators developed by Google | very large-scale training or inference should run efficiently |
| GPU | – | Infrastructure | AI practitioners | General-purpose parallel processors | flexible acceleration is needed |
| AI Hypercomputer | – | Infrastructure | AI practitioners | Integrated supercomputing architecture | gen AI workloads run at the largest scale |
| Gemini Nano + LiteRT | TensorFlow Lite | Infrastructure/Models (edge) | Mobile and edge developers | On-device AI, offline, private, real-time | latency, offline operation or on-device privacy are critical |
| IAM | – | Infrastructure (security) | Admins | Access control (AuthN + AuthZ) | you need to control who is allowed to do what |
| Security Command Center | – | Infrastructure (security) | Security teams | Security posture, vulnerabilities | you need an overview of risks |
| Sensitive Data Protection | Cloud DLP | Infrastructure (security) | Data and security teams | Detect and de-identify PII | sensitive data must be protected before training or prompting |
| Mandiant / Google Threat Intelligence | – | Services | Security teams | Threat intel, incident expertise | proactive defense and expert help are needed |
| SAIF (framework, not a product) | – | Governance | Leadership, security | Best practices for secure AI | a framework for AI security is wanted |
The 6 most important distinctions at a glance
| Commonly confused pair | Distinguishing feature |
|---|---|
| Gemini Enterprise vs. Gemini Enterprise Agent Platform | App for employees vs. developer platform (formerly Vertex AI) |
| Gemini Enterprise vs. Gemini Enterprise for CX | Internal employees vs. external customers (contact center) |
| Agent Studio vs. Google AI Studio | Professionals, production, cloud governance, paid vs. beginners, prototype, free |
| Agent Assist vs. Conversational Agents | Helps the human in the conversation vs. replaces them as a bot |
| Conversational Insights vs. Agent Assist | Analysis of all conversations afterwards vs. help live in a single conversation |
| Gemini app vs. Gemini for Workspace vs. Gemini Notebook | General chat vs. AI inside Gmail, Docs, etc. vs. only your own sources |
Sources
Exam Guide §1.3, §1.4, §2.1–2.5, §4.2 · Study Guide pp. 4–7, p. 10 · cloud.google.com/gemini-enterprise-agent-platform/generative-ai/pricing · deepmind.google/models/lyria · docs.cloud.google.com/speech-to-text/docs/models/chirp-3 · saif.google · docs.cloud.google.com/gemini-enterprise-agent-platform/release-notes (naming table dated 2026-04-22) · docs.cloud.google.com/gemini-enterprise-agent-platform/overview · blog.google/innovation-and-ai/products/gemini-notebook/notebooklm-gemini-notebook · cloud.google.com/solutions/customer-engagement-ai
Note: The former names Duet AI, MakerSuite, PaLM API, TensorFlow Lite and Google AI Pro do not come from Exam Guide/Study Guide but are own classification. Cloud DLP → Sensitive Data Protection is documented at docs.cloud.google.com/sensitive-data-protection/docs/sensitive-data-protection-overview, the Vertex AI renamings in the release notes (naming table dated 2026-04-22).
Read next: Section 2: Google Cloud's gen AI offerings · Glossary · Google Cloud Generative AI Leader