Reproducible example

Classify customer feedback from a CSV

Use a Choice question to route each feedback row into one of six distinct topics while keeping probabilities for later review.

Sample CSV

The downloadable dataset contains fictional feedback only. It has four columns: feedback_id, feedback_text, channel, and created_at.

Question configuration

Type: choice
Name: feedback_topic
Instructions: Which topic best describes this customer feedback?

billing       Payments, invoices, charges, or refunds.
product_bug   Product behavior that appears broken or incorrect.
account       Login, permissions, profile, or account access.
feature       A request for new or changed product capability.
cancellation  Requests to cancel, close, or stop a subscription.
other         Feedback outside the categories above.

Illustrative output shape

FeedbackChoiceConfidence
I was charged twice this month.billingHigh
Export is blank in Safari.product_bugHigh
Please add an API usage dashboard.featureHigh
I cannot invite another admin.accountMedium

These labels describe the expected demonstration behavior, not a published accuracy benchmark. Actual model output can vary, and thresholds should be evaluated on representative labeled data.

What to inspect before using the result

Sources and verification

Product facts on this page were checked against the following primary sources on October 11, 2026.

Use your data

Replace the sample with your feedback

Keep the configuration, select your feedback column, and preview the first 20 real rows.

Try my customer feedback CSV