> ## Documentation Index
> Fetch the complete documentation index at: https://docs.useinari.com/llms.txt
> Use this file to discover all available pages before exploring further.

# CSV

> Ingest new feedback through CSV uploads

## Add New Feedback via CSV

You can add any types of feedback data through `Uploaded CSVs` in the `Sources` page, accessible from the nav bar.

Every feedback requires 3 fields while uploading:

* **Customer Name**: the full name of the customer that submitted the feedback
* **Date Received**: the date that the feedback was sent by the customer
* **Feedback**: the actual content of the feedback aggregated in one column

There's a variety of other fields like Company Name, NPS, CSAT, and unique ID's - pull those together and map them into Inari whenever possible since this further enriches the quality of insights and metrics generated but they are **optional**.

## Space Triage

You can assign or `Triage` a piece of feedback into a specific `Space` when adding a new data source or updating the `Space` value in the feedback's detail panel.

Feedbacks that are `Triaged` into a `Space` will be used as context when generating new insights and issues in that `Space`. Existing insights and issues in a given space will also be auto-linked to feedback triaged into that space.

This allows for more concentrated and higher-quality insights and issues to be generated since triaging allows for higher-quality clusters to arise.

## Summaries

You can optionally create a `Summary` - this will create a brief summary of the top takeaways and useful quotes from the full customer interaction.

We recommend only setting this on for long documents or customer conversations with long back and forth messages - the summary will pull the top 4 takeaways and snippets. For shorter documents, summaries will often be duplicative and run the risk of hallucinating quotes so we strongly encourage not to use them for short feedback.

## Filter Spam

You can optionally pre-filter all feedback to check whether the feedback seems like `Spam` and is not product feedback.

This is most useful when analyzing feedback from sources similar to Intercom or Zendesk where high volumes of emails and support conversations aren't necessarily product feedback and should be removed.

We recommend only setting this on for data sources riddled with marketing or spam messages since the AI can make mistakes and miscategorize feedback.