CSV Upload

Lexsis AI supports uploading feedback data sets via CSV or Excel files. This allows you to bulk import historical feedback or sync data from systems that don't have direct API integrations.

Overview

CSV upload enables you to:

• Upload large batches of feedback at once

• Import historical feedback data

• Sync data from systems without API integrations

• Include rich metadata for better classification

File Format Requirements

• File types: CSV (.csv)

• Encoding: UTF-8 (Latin-1 is also accepted as a fallback)

• Delimiter: Comma (,)

• Header row: First row must contain column names

• Maximum file size: 50 MB

• Maximum rows: 25,000 rows per file

• Maximum columns: 50 columns per file

Required Columns

Only the following column is required:

Column Name	Type	Description
raw_content	string	The raw text feedback that you'd like Lexsis AI to process

Tip: You can also use the alias content instead of raw_content.

Example (Minimum Required)

raw_content
"The app crashes when I try to upload photos"
"Love the new dashboard design!"
"Customer support is slow to respond"

Recommended Columns

The following columns are highly recommended for best results. If not provided, they will be auto-generated:

Column Name	Type	Description	Default if missing
submitted_at	datetime	The date and time that the feedback was submitted. Accepts ISO 8601 format (YYYY-MM-DDTHH:MM:SSZ) or date format (YYYY-MM-DD)	Current date/time at upload
source_identifier	string	A unique identifier for the feedback submission (e.g., review ID, ticket ID). Must be unique within your tenant	Auto-generated unique identifier

Example with Recommended Columns

raw_content,submitted_at,source_identifier
"The app crashes when I try to upload photos",2024-01-15T10:30:00Z,review-12345
"Love the new dashboard design!",2024-01-15T11:00:00Z,review-12346

Column Aliases

For convenience, the following aliases are supported. You can use either the alias or the standard column name in your CSV header:

Alias	Maps to
content	raw_content
date	submitted_at
id	source_identifier

Example Using Aliases

content,date,id
"The app crashes when I try to upload photos",2024-01-15T10:30:00Z,review-12345
"Love the new dashboard design!",2024-01-15T11:00:00Z,review-12346

Optional Columns - Core Submission Fields

Column Name	Type	Description
source_id	UUID	UUID of the source integration. If not provided, a default source will be used
title	string	Optional title for the feedback submission

Optional Columns - Metadata Fields

All metadata columns map directly to the related_metadata object in the submission. Column names must match the exact field names from the data model (case-sensitive).

Customer/User Information

Column Name	Type	Description
customer_id	string	Unique identifier for the customer/user
customer_name	string	Customer/user name
customer_email	string	Customer email address
customer_segment	string	Customer segment. Must be one of: Enterprise, SMB, Free, Trial
customer_value	float	Monetary value of customer (MRR, LTV, etc.). Used for priority calculation

Business/Organization Information

Column Name	Type	Description
business_name	string	Business/organization name
business_id	string	Unique identifier for business/organization

Platform/Product Information

Column Name	Type	Description
platform	string	Platform where feedback was submitted. Must be one of: iOS, Android, Web, Desktop, Mobile, Tablet
app_version	string	Application version (e.g., 4.94.0)
product_version	string	Product version
review_created_version	string	App version when review was created

Customer Journey Information

Column Name	Type	Description
account_age_days	integer	Number of days since account creation. Used for journey stage inference
usage_frequency	string	Frequency of product usage. Must be one of: Daily, Weekly, Monthly, Rarely, Never
subscription_status	string	Subscription status. Must be one of: Active, Trial, Expired, Cancelled

Rating/Score Information

Column Name	Type	Description
score	integer	Rating score (1-5 stars)
thumbs_up_count	integer	Number of thumbs up/helpful votes

Location/Geography

Column Name	Type	Description
location	string	Geographic location (country, region, etc.)
country	string	Country code (ISO 3166-1 alpha-2, e.g., us, gb)
language	string	Language code (ISO 639-1, e.g., en, es)

Response/Reply Information

Column Name	Type	Description
reply_content	string	Response/reply content from business
replied_at	datetime	Timestamp when reply was posted (ISO 8601 format)

Custom Fields

If your data set includes custom fields that don't map to the standard metadata fields, you can include them in your CSV. The uploader allows you to:

1. Select any additional columns to bring in as custom fields

2. Custom fields are preserved for reference and can be used in filtering and analysis

Example CSV Files

Minimal Example (Required Column Only)

raw_content
"The app crashes when I try to upload photos"
"Love the new dashboard design!"
"Customer support is slow to respond"

Example with Recommended Columns

raw_content,submitted_at,source_identifier
"The app crashes when I try to upload photos",2024-01-15T10:30:00Z,review-12345
"Love the new dashboard design!",2024-01-15T11:00:00Z,review-12346
"Customer support is slow to respond",2024-01-15T12:00:00Z,review-12347

Example with Aliases

content,date,id
"The app crashes when I try to upload photos",2024-01-15T10:30:00Z,review-12345
"Love the new dashboard design!",2024-01-15T11:00:00Z,review-12346
"Customer support is slow to respond",2024-01-15T12:00:00Z,review-12347

Example with Metadata

raw_content,submitted_at,source_identifier,customer_name,platform,app_version,score,customer_segment,customer_value
"The app crashes when I try to upload photos",2024-01-15T10:30:00Z,review-12345,John Doe,iOS,2.1.0,2,Enterprise,5000.00
"Love the new dashboard design!",2024-01-15T11:00:00Z,review-12346,Jane Smith,Web,2.1.0,5,SMB,500.00
"Customer support is slow to respond",2024-01-15T12:00:00Z,review-12347,Bob Wilson,Android,2.0.5,3,Free,0.00

Example with Journey Information

raw_content,submitted_at,source_identifier,customer_name,account_age_days,usage_frequency,subscription_status,country
"Having trouble logging in",2024-01-15T10:30:00Z,review-12345,John Doe,45,Daily,Active,us
"Feature request: dark mode",2024-01-15T11:00:00Z,review-12346,Jane Smith,120,Weekly,Active,gb
"Billing question",2024-01-15T12:00:00Z,review-12347,Bob Wilson,30,Monthly,Trial,ca

Example with Custom Fields

raw_content,submitted_at,source_identifier,customer_name,score,department,internal_priority
"Bug report: upload feature broken",2024-01-15T10:30:00Z,review-12345,John Doe,1,Engineering,High
"Feature request: export data",2024-01-15T11:00:00Z,review-12346,Jane Smith,4,Product,Medium
"Question about pricing",2024-01-15T12:00:00Z,review-12347,Bob Wilson,3,Sales,Low

Column Mapping

Core Fields

CSV columns map directly to submission fields:

• raw_content (or content) → raw_content (submission field)

• submitted_at (or date) → submitted_at (submission field, auto-generated if missing)

• source_identifier (or id) → source_identifier (submission field, auto-generated if missing)

• source_id → source_id (submission field, optional)

Important Notes

Column Name Requirements

• Case-sensitive: Column names must match exactly (e.g., customer_id not Customer_ID or customerId)

• Exact spelling: Column names must match the data model field names exactly (or use a supported alias)

• No spaces: Use underscores, not spaces (e.g., customer_name not customer name)

• No duplicates: Each column name must appear only once in the header row

Data Type Validation

• Dates: Use ISO 8601 format (YYYY-MM-DDTHH:MM:SSZ) or date format (YYYY-MM-DD)

• Integers: Must be valid integers (e.g., score must be 1-5)

• Floats: Must be valid decimal numbers (e.g., customer_value must be a number)

• Enums: Must match exact values (e.g., platform must be one of the allowed values)

Unique Identifiers

The source_identifier must be unique within your tenant. Duplicate identifiers will cause those rows to be skipped.

Upload Limits

• File size: Maximum 50 MB per file

• Row limit: Maximum 25,000 rows per file

• Column limit: Maximum 50 columns per file

• Processing time: Typically 1-5 minutes per 1,000 rows, depending on metadata complexity

Processing

Once uploaded successfully:

1. Validation: CSV is validated for format and data integrity

2. Import: Data is imported into Lexsis AI

3. Processing: Submissions are automatically processed through the pipeline (if auto-processing is enabled):

   • Segmentation

   • Classification

   • Bucketing

4. Completion: You'll receive a notification when processing is complete

Next Steps

• Submit Feedback via REST API - Submit feedback programmatically

• API Overview - Explore other API endpoints

• Getting Started Guide - Set up your account