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

# Data format: AI coding for open-ended responses

> Automatically code and categorize open-ended survey responses at scale using AI-powered topic classification — from upload to Excel export.

Pulse Classifier is Smartinterview's AI tool for coding open-ended survey responses. Instead of manually reading and tagging thousands of verbatim answers, you upload your data, let the AI propose a topic code plan, review and refine it, then run classification across your full dataset. Results export to Excel with structured topic sheets ready for analysis.

Use Pulse Classifier when you have open-ended survey questions — either from an Excel or SPSS file you already own, or from a Smartinterview survey you ran — and need to turn free-text answers into structured, countable topic codes.

## End-to-end workflow

<Steps>
  <Step title="Upload your data">
    Upload an Excel (`.xlsx` / `.xls`) or SPSS (`.sav`) file, or connect directly to an existing Smartinterview survey question. You then map which sheet, respondent ID column, and response text column to use.

    [Upload data →](/pulse-classifier/upload-data)
  </Step>

  <Step title="Generate topics">
    Click **Generate Topics** to let the AI read a sample of your responses and propose a topic code plan. Review the suggested topics, rename or delete any that don't fit, add your own, and optionally expand any topic into L1/L2 sub-topics.

    [Define and manage topics →](/pulse-classifier/topics)
  </Step>

  <Step title="Configure and classify">
    Set classification depth, tolerance, and maximum codes per response. Run a preview classification on a small sample to validate topic quality, then launch the full classification job and monitor progress in real time.

    [Classify responses →](/pulse-classifier/classify)
  </Step>

  <Step title="Export and analyze">
    Download results as Excel — your original data plus topic assignment columns, a **Topics** summary sheet, and a **Top Topics** sheet. Explore topic distributions, co-occurrence heatmaps, and sentiment summaries directly in the platform.

    [Export results →](/pulse-classifier/export)
  </Step>
</Steps>

## When to use Pulse Classifier

Pulse Classifier is best suited for:

* **Survey open-ends** — Net Promoter Score verbatims, satisfaction open-ends, "Why did you choose us?" questions, and similar free-text fields collected in Excel or via Smartinterview.
* **Large response volumes** — Hundreds to tens of thousands of responses where manual coding would take days.
* **Iterative coding projects** — When you need to refine topic definitions over multiple rounds or reuse a topic frame across survey waves.
* **Multi-column files** — Files with several open-ended columns that each need their own classification plan.

<Tip>
  If your responses are in a language other than English, Pulse Classifier handles multilingual text automatically. You can write your topic labels in any language and the AI will match responses accordingly.
</Tip>

## Key concepts

| Term          | Definition                                                                                                                      |
| ------------- | ------------------------------------------------------------------------------------------------------------------------------- |
| **Topic**     | A thematic code assigned to a response (e.g., "Price", "Customer service").                                                     |
| **L1 / L2**   | Classification depth. L1 = top-level topics only. L2 = L1 parent topics with sub-topics beneath them.                           |
| **Tolerance** | Controls how strictly the AI applies topic rules. Lower values are more conservative; higher values are more inclusive.         |
| **Consensus** | The AI runs multiple classification passes per response and keeps only topics that appear in enough passes, improving accuracy. |
| **FilesQO**   | The primary output sheet in the Excel export — one row per response with all assigned topic columns.                            |