Stay organized with collections
Save and categorize content based on your preferences.
The Agent Assist Article
Suggestion and FAQ
Assist features follow a conversation between a human
agent and an end-user and provide the human agent with relevant document
suggestions. Both features make suggestions based on
knowledge bases, collections of documents that you upload to
Agent Assist. These documents are called knowledge documents and can be
either articles (for use with Article Suggestion) or FAQ documents (for use with
FAQ Assist). See the knowledge documents
documentation for best practices
information and the data overview page for more
details on the types of data that can be uploaded to Agent Assist. This
tutorial walks you through the steps required to create a knowledge base and add
documents to it using the Agent Assist Console.
Before you begin
Follow the Dialogflow setup
instructions to enable Dialogflow on a Google Cloud Platform project.
We recommend that you read the Agent Assist basics
page before starting this tutorial.
If you are creating a knowledge base for use with Article
Suggestion, make sure that your
knowledge documents are in one of two formats:
If your FAQ document is in .csv format, it must contain two columns. FAQ
questions must be listed in the first column, and the answers to each
question must be listed in the second column. Each FAQ question and its
associated answer is called an FAQ pair. Make sure that the .csv file
does not contain a header row.
Console. Select your Google Cloud project, then click the
Data menu option on the far left margin of the page. The Data menu
displays all of your data. There are two tabs, one each for conversation
datasets and
knowledge bases.
Click knowledge bases, then the +Create new button at the top of the
knowledge bases page.
In the menu that pops up enter a name for the knowledge base and choose a
language.
You will see a list (currently empty) of all documents that are part of this
knowledge base. To add a document, click +Create New.
Under choose a knowledge type select FAQ if you're uploading FAQ
documents for use with FAQ Assist, or
article suggestion if you're uploading articles for use with Article
Suggestion.
Under choose a file source, choose the location of the document you're
adding. See the before you
begin section for
details on supported locations and file formats.
[[["Easy to understand","easyToUnderstand","thumb-up"],["Solved my problem","solvedMyProblem","thumb-up"],["Other","otherUp","thumb-up"]],[["Hard to understand","hardToUnderstand","thumb-down"],["Incorrect information or sample code","incorrectInformationOrSampleCode","thumb-down"],["Missing the information/samples I need","missingTheInformationSamplesINeed","thumb-down"],["Other","otherDown","thumb-down"]],["Last updated 2025-08-28 UTC."],[[["\u003cp\u003eAgent Assist's Article Suggestion and FAQ Assist provide human agents with relevant document suggestions during conversations with end-users, based on uploaded collections of documents called knowledge bases.\u003c/p\u003e\n"],["\u003cp\u003eKnowledge documents, which can be either articles or FAQ documents, must be in specific formats, such as a public URL or a file stored in a Cloud Storage bucket, depending on whether they are for Article Suggestion or FAQ Assist.\u003c/p\u003e\n"],["\u003cp\u003eCreating a knowledge base in the Agent Assist Console involves selecting a project, navigating to the Data menu, creating a new knowledge base, choosing a knowledge type (FAQ or article suggestion), and adding documents one at a time.\u003c/p\u003e\n"],["\u003cp\u003eIf uploading FAQ documents in \u003ccode\u003e.csv\u003c/code\u003e format, the file must have two columns containing the FAQ questions and answers, respectively, without a header row.\u003c/p\u003e\n"],["\u003cp\u003eWhile the Agent Assist Console requires adding documents one at a time, the Agent Assist API is recommended for uploading large quantities of documents, with a suggestion of at least 50 documents per knowledge base.\u003c/p\u003e\n"]]],[],null,["# Create a knowledge base\n\nThe Agent Assist [Article\nSuggestion](/agent-assist/docs/article-suggestion) and [FAQ\nAssist](/agent-assist/docs/faq) features follow a conversation between a human\nagent and an end-user and provide the human agent with relevant document\nsuggestions. Both features make suggestions based on\n*knowledge bases* , collections of documents that you upload to\nAgent Assist. These documents are called *knowledge documents* and can be\neither articles (for use with Article Suggestion) or FAQ documents (for use with\nFAQ Assist). See the [knowledge documents\ndocumentation](/agent-assist/docs/knowledge-documents) for best practices\ninformation and the [data overview page](/agent-assist/docs/data) for more\ndetails on the types of data that can be uploaded to Agent Assist. This\ntutorial walks you through the steps required to create a knowledge base and add\ndocuments to it using the Agent Assist Console.\n\nBefore you begin\n----------------\n\n1. Follow the [Dialogflow setup](https://cloud.google.com/dialogflow/es/docs/quick/setup?hl=en) instructions to enable Dialogflow on a Google Cloud Platform project.\n2. We recommend that you read the Agent Assist [basics\n page](/agent-assist/docs/basics) before starting this tutorial.\n3. If you are creating a knowledge base for use with [Article\n Suggestion](/agent-assist/docs/article-suggestion), make sure that your knowledge documents are in one of two formats:\n - A public URL.\n - A file stored in a [Cloud Storage\n bucket](/storage/docs/creating-buckets).\n4. If you are creating a knowledge base for use with [FAQ\n Assist](/agent-assist/docs/faq), make sure that your FAQ documents are in\n one of three formats:\n\n - A public URL.\n - A `csv` file saved in a [Cloud Storage\n bucket](/storage/docs/creating-buckets).\n - The contents of a `csv` file.\n\n If your FAQ document is in `.csv` format, it must contain two columns. FAQ\n questions must be listed in the first column, and the answers to each\n question must be listed in the second column. Each FAQ question and its\n associated answer is called an **FAQ pair** . Make sure that the `.csv` file\n does not contain a header row.\n\nCreate a knowledge base\n-----------------------\n\n1. Navigate to the [Agent Assist](https://agentassist.cloud.google.com)\n\n [](https://agentassist.cloud.google.com)\n\n [Console](https://agentassist.cloud.google.com). Select your Google Cloud project, then click the\n **Data** menu option on the far left margin of the page. The **Data** menu\n displays all of your data. There are two tabs, one each for **conversation\n datasets** and\n **knowledge bases**.\n2. Click **knowledge bases** , then the **+Create new** button at the top of the\n knowledge bases page.\n\n3. In the menu that pops up enter a name for the knowledge base and choose a\n language.\n\n4. You will see a list (currently empty) of all documents that are part of this\n knowledge base. To add a document, click **+Create New**.\n\n5. Under **choose a knowledge type** select **FAQ** if you're uploading FAQ\n documents for use with [FAQ Assist](/agent-assist/docs/faq), or\n **article suggestion** if you're uploading articles for use with [Article\n Suggestion](/agent-assist/docs/article-suggestion).\n\n6. Under **choose a file source** , choose the location of the document you're\n adding. See the [before you\n begin](/agent-assist/docs/knowledge-base#before_you_begin) section for\n details on supported locations and file formats.\n\n7. Click **Create** to add the document.\n\n | **Note:** When you use the Agent Assist Console you must add documents to a knowledge base one at a time. If you want to upload a large number of documents, we recommend that you use the [Agent Assist API](/agent-assist/docs/article-suggestion). We recommend that each knowledge base contain at least 50 documents.\n\nWhat's next\n-----------\n\n[Create a conversation profile](/agent-assist/docs/conversation-profile) using\nthe Agent Assist Console."]]
