Google Cloud sells two families of product for the moment a customer wants something. Agent Search covers finding, across the websites, records and documents an organisation owns. The Customer Engagement Suite covers the conversation, from an automated chat through to the telephony a contact centre runs on. Both serve the same kind of business, one whose customers arrive with a question and expect an answer without waiting for it.
The sections below take Agent Search first, then Google Search beside it, then the four parts of the Customer Engagement Suite, and close on the detail that tells one product from another in an exam question.
Agent Search
Agent Search applies Google's search technology to content an organisation owns. Until Google renamed the platform around it, the product was called Vertex AI Search, so a candidate meets both names in the study material and answers with the current one. It indexes a website, structured records such as a product catalogue and unstructured documents, and it returns results with the semantic understanding people expect of a search engine. The business cases are a site search that returns what a visitor meant, a product finder that works from a description and internal search across the documents a company has accumulated.
Agent Search also does the retrieval behind grounding. A grounded model on Google Cloud usually takes its supporting passages from an Agent Search index, which is why the product appears again in the material on grounding and retrieval.
Google Search
Google Search sits beside Agent Search in the exam guide because it covers the opposite corpus. Where a model needs current public information, grounding with Google Search supplies it. The two products appear together because each gives a model something outside itself to answer from, and the question that separates them is whose content the answer comes from.
The Customer Engagement Suite
The Customer Engagement Suite is Google Cloud's contact centre product, and it comes in four parts. A question in this section describes a situation in a support operation and asks which of the four it calls for, so what matters about each part is who it helps and when.
Conversational Agents are the customer facing automation. A chat or voice agent works out what a customer wants and completes it, answering from the organisation's own content and calling its systems. This is the part of the suite that handles a whole conversation with nobody else in it.
Agent Assist supports the human agent while a conversation is running. It suggests replies, surfaces the article the agent needs and writes the summary at the end, so nobody types up the call afterwards. The customer never sees it, and what it returns to the business is shorter handling time and more consistent answers.
Conversational Insights analyses conversations once they have finished, across the whole volume of them. It reports why people are contacting the business, which topics drive the longest calls and where conversations break down. The output informs a roadmap.
Contact Center as a Service is the contact centre platform itself, delivered as a managed service. It supplies the telephony and the routing that the other three parts run on.
Which product a described situation needs
The detail that separates one product from another is who each one helps. Conversational Agents help the customer directly, with no person in the conversation. Agent Assist helps the person taking the call, while the call is happening. Conversational Insights helps whoever decides what to change, by reporting what happened across thousands of conversations. Contact Center as a Service is the platform underneath all three. A question about finding something points away from the suite altogether, at Agent Search over an organisation's own content or at Google Search over the public web.