Skip to main content
The FAQ connector lets you upload frequently asked questions and their answers to a knowledge base. The uploaded FAQ content is indexed and used by the knowledge base to retrieve relevant information and generate answers to user questions. You can upload FAQ data in JSON or CSV format and configure how each field is used for indexing, answering, and citations. The connector supports mapping FAQ fields, adding custom fields, configuring citations, and choosing how FAQ records are chunked for indexing.

​
Step 1: Add an FAQ source

  1. Navigate to the knowledge base.
  2. Select FAQ as the source.
  3. Enter a name for the FAQ source.
  4. Upload one or more JSON or CSV files.
  5. You can also add FAQs manually instead of uploading a file.
The FAQ source provides sample JSON and CSV files that you can download and use as templates.

​
Step 2: Prepare the FAQ file

Required fields All the other optional fields present in the file are also uploaded and can be mapped to index fields.

​
Step 3: Map file fields

After you upload a file, the connector displays the fields detected in the file. The connector automatically matches file keys to FAQ fields when their names match. Fields that do not match can be mapped manually. For each field, select or review the corresponding FAQ field from the Maps to dropdown. For example, map the file keys to the corresponding FAQ fields as shown below.
Field usage

​
Step 4: Add custom fields

If your FAQ file contains information that is not covered by the default fields, you can create custom fields.
  1. Select Add custom field.
  2. Define the custom field.
  3. Map the corresponding field from your uploaded file.
  4. Select whether the field should be embedded, included in the answer, or excluded.
Up to 10 custom string fields can be used.

​
Step 5: Configure chunking

Under Advanced settings, select how FAQ records should be chunked for indexing.
  • Standard - by FAQ: Creates one chunk per FAQ. This is the default option and provides faster indexing.
  • Precise - by question: Creates one chunk per question, including the answer to the question. This option provides more precise question-level indexing but takes longer to index.

​
Step 6: Save the FAQ source

After uploading the file and configuring field mappings and chunking:
  1. Review the field mappings.
  2. Verify that Question and Answer are mapped correctly.
  3. Verify the usage configured for each field.
  4. Select the required chunking strategy.
  5. Click Save.
The connector processes the uploaded FAQs. Once the FAQ source is indexed and available in the knowledge base, an agent can use it to retrieve relevant FAQs and answer user questions.

​
Points to Note

  • Use JSON or CSV format for the FAQ data file.
  • File size limit: 100 MB per file.
  • Custom field limit: 10 custom fields.