GCP Professional Cloud Architect Practice Question

A global retail company plans to pilot a generative-AI concierge that can: (1) answer detailed product questions using hundreds of existing PDF manuals, (2) retrieve order status from an internal REST service, and (3) hand off the conversation to a live agent. Non-technical product owners must be able to visually design and iterate on the dialog, and the assistant should be published to both the public website chat widget and a phone voice channel with as little engineering effort as possible. Which Google Cloud capability best meets these needs?

Prototype the assistant in Vertex AI Studio notebooks and deploy the code as Cloud Run services behind an API Gateway.
Build a bespoke chatbot on Google Kubernetes Engine using open-source large language models and expose it through Apigee APIs.
Create the concierge in Dialogflow CX using generative knowledge connectors and its built-in web and telephony integrations, with a Cloud Function webhook for order lookup.
Use Agent Builder with knowledge connectors and rely on custom integrations for web and voice channels.

Report Issue

Answer Description

Dialogflow CX supplies a drag-and-drop conversation builder suitable for non-developers, supports Generative AI and knowledge connectors that ingest PDF documents for retrieval-augmented responses, and includes out-of-the-box integrations for web chat and a managed telephony gateway. Fulfillment can call an internal REST endpoint through a lightweight Cloud Function, keeping engineering effort minimal while letting product managers own the dialog. Agent Builder lacks turnkey voice deployment, while the Vertex AI Studio and GKE options require substantially more custom coding and infrastructure, making them less aligned with the stated constraints.

Ask Bash

Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.