Vertex AI Agent Builder is GCP's Gen AI agent-building platform, bundling Conversational Agents (the Dialogflow CX successor), Vertex AI Search, Tool Use, and Grounding into a single offering. It lets you build production-grade AI agents with no-code or low-code, and it has been Google's strategic agent service since 2024.
| Type | Description | Use Case |
|---|---|---|
| Deterministic Agent | Traditional intent / flow definitions | Fixed FAQs, business procedures |
| Generative Agent (Playbook) | Define flows in natural language | Flexible dialogue, general advisory |
| Hybrid | Combines both | Business tasks + small talk |
| Variant | Data Source | Use Case |
|---|---|---|
| Search for Web sites | Public URLs / sitemap | On-site search |
| Search for Documents | GCS / BigQuery / Cloud SQL | Internal document search |
| Search for Healthcare | FHIR / healthcare data | Healthcare |
| Search for Media | Video / audio | Media companies |
| Recommendations | Products / content | Recommendations |
# Goal
Help the user check the status of their delivery
# Instructions
1. Confirm the order number
2. Call ${TOOL: get_order_status} to retrieve the status
3. If the order is in transit, return the carrier and tracking URL
4. If the order is delayed, route the user to ${TOOL: contact_support}
# Examples
User: I want to check on my order
Agent: Could you share your order number?
User: ORD-12345
Agent: Looking it up... ORD-12345 is in transit with Yamato Transport, tracking number ...openapi: 3.0.0
info:
title: Order API
version: 1.0.0
paths:
/orders/{order_id}/status:
get:
operationId: get_order_status
parameters:
- name: order_id
in: path
required: true
schema: {type: string}
responses:
'200':
description: Order status
content:
application/json:
schema:
properties:
status: {type: string}
tracking_url: {type: string}| Item | Price |
|---|---|
| Conversation Search Standard | $4 per 1,000 queries |
| Conversation Search Enterprise | $25 per 1,000 queries |
| Vertex AI Search (search only) | $2 per 1,000 queries |
| Generative Agent | Sessions + Gemini tokens |
| Tool invocations | Priced by the underlying API |
| Item | Agent Builder | ChatGPT GPTs / Assistants | Copilot Studio | Amazon Bedrock Agents |
|---|---|---|---|---|
| Underlying LLM | Gemini | GPT-4o | GPT-4 / Phi | Claude / Llama |
| No-code | Excellent | Excellent | Excellent | Good |
| Multimodal | Excellent | Excellent | Good | Good |
| Internal data integration | Vertex AI Search | File Search | M365 + SharePoint | Bedrock KB |
| Pricing | $4-25 per 1k queries | Subscription + tokens | From $200/user/month | Tokens |
What is Vertex AI Agent Builder?
An integrated no-code / low-code platform for building Gen AI agents (Conversational Agent + Search + Tool Use). It is the successor to Dialogflow CX, rebuilt on top of Gemini.
How does it differ from Conversational Agents (formerly Dialogflow CX)?
Under the hood they are the same service, but rebranded as Agent Builder and rearchitected around Gen AI, Generative Playbook, and Tool integration. It is a completely separate product from the legacy Dialogflow ES.
What is Vertex AI Search?
A search service from the Discovery Engine family. It lets you spin up LLM + RAG search over internal documents, websites, or structured data in a single click.
How is it priced?
Conversation Search: $4 per 1,000 queries (Standard) / $25 (Enterprise). Agent: sessions + LLM tokens. Vertex AI Search: $2 per 1,000 queries (search only).
How do you implement Tool Use?
Register an OpenAPI spec as a Tool, and the Agent will invoke the API as needed. It works with Cloud Functions, API Gateway, or any REST API.
What can Grounding do?
You can combine internal data (via Vertex AI Search), fresh web information (via Google Search), and real-time data (via your own webhooks) as grounding sources for responses.
How does it compare to ChatGPT GPTs / Microsoft Copilot Studio?
GPTs is OpenAI-only and Copilot Studio is tightly tied to M365. Agent Builder's strengths are Gemini, multimodal support, and deep integration with GCP data services.
What is a Playbook?
A Generative Agent feature for defining flows in natural language. You write rules like "when the user asks X, do Y" in Markdown-style prose, defining agent behavior without writing code.
Related Articles: Gen AI Agents
Vertex AI Fundamentals for GCP Certs (2026)
Vertex AI basics every cert candidate needs — Workbench, Pipelines, Model Garden, Endpoints.
GAIL Generative AI Leader: Complete Exam Guide (2026)
Pass the Generative AI Leader exam — Gemini, Vertex AI, Workspace AI for business leaders.
PMLE GenAI Implementation: Vertex AI + RAG (2026)
GenAI implementation patterns for the PMLE exam — embeddings, vector search, RAG, evaluation.
Vertex AI Feature Store Complete Guide (2026)
Feature Store fundamentals — online/offline serving, point-in-time, common ML feature patterns.
* Google Cloud and Vertex AI are trademarks of Google LLC. For the latest details, see the official Agent Builder page.
Practice with certification-focused question sets
See the GCP exam prep pageNicheeLab Editorial Team
NicheeLab editorial team focused on data engineering and cloud certification learning. Content is structured around practical study needs and official exam domains.
Google Cloud Certification Roadmap (2026)
Choose your GCP certification path — Foundational, Associate...
CDL Cloud Digital Leader: Complete Exam Guide (2026)
Pass the Cloud Digital Leader exam — cloud business value, G...
GAIL Generative AI Leader: Complete Exam Guide (2026)
Pass the Generative AI Leader exam — Gemini, Vertex AI, Work...
Vertex AI Fundamentals for GCP Certs (2026)
Vertex AI basics every cert candidate needs — Workbench, Pip...
Associate Cloud Engineer (ACE): Complete Guide (2026)
Pass the Associate Cloud Engineer exam — Console, gcloud, pr...