> ## 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.
# Question types in the Survey Builder
> Learn about every question type in the Survey Builder: text, choice, scale, vocal, AI follow-up, and image — with guidance on when to use each.
The **Survey Builder** provides nine question types that cover the full range of qualitative and quantitative data collection. Each type is optimised for a different kind of respondent interaction, from a quick single-choice answer to an AI-driven conversational follow-up. Choose the type that best matches the data you need, and combine multiple types within a single survey.
## Text questions
**What it is:** A single-line text input that accepts a brief free-form answer.
**When to use it:** Collect a name, a short label, a numeric identifier, or any open-ended answer you expect to be one or two sentences at most.
**How to add it:** Click **Add question**, then select **Text** from the question type picker. The input renders as a standard single-line field for the respondent.
**What it is:** A multi-line text area that gives respondents space to write a detailed answer.
**When to use it:** Ask open-ended opinion questions, gather verbatim feedback, or collect anything that typically requires a paragraph or more. Long-text answers are well-suited for classification with **Pulse Classifier** after the survey ends.
**How to add it:** Select **Text** and enable the multi-line option in the question editor. The input expands to a resizable text area for respondents.
## Choice questions
**What it is:** A list of options where the respondent can choose exactly one answer.
**When to use it:** Ask for a single definitive answer — preferred product, age bracket, satisfaction rating as a category, or any situation where picking multiple options would not make sense. Radio group questions can also serve as the basis for [quota segmentation](/survey-builder/quotas).
**How to add it:** Select **Radio Group** from the question type picker, then add or edit the answer choices in the question editor.
**What it is:** A list of options where the respondent can select any number of answers.
**When to use it:** Ask which products a respondent has used, which topics interest them, or any scenario where multiple answers are valid simultaneously.
**How to add it:** Select **Checkbox Group** from the question type picker, then configure the answer choices.
**What it is:** A collapsed select menu that shows all options when the respondent clicks it.
**When to use it:** Use a dropdown when you have many options (more than five or six) and want to keep the question visually compact. Like radio groups, dropdown questions can be used to define quota segmentation.
**How to add it:** Select **Dropdown** from the question type picker and add your options.
## Scale and rating questions
**What it is:** A numeric slider that lets respondents pick a value along a defined range. The default range is 1–5, but you can adjust the minimum and maximum values in the question editor.
**When to use it:** Collect numeric ratings, satisfaction scores, likelihood-to-recommend questions (NPS-style), or any measurement where position on a scale matters. Scale answers are numeric and can be used in conditions with `>`, `<`, `>=`, and `<=` operators.
**How to add it:** Select **Slider** from the question type picker. Set the minimum and maximum values to match your intended scale.
Pair a scale question with a text question immediately after it if you want respondents to explain their rating. Use [skip logic](/survey-builder/skip-logic) to show the follow-up only when the score is below a threshold (for example, show "What could we improve?" only when the rating is ≤ 2).
## Vocal questions
**What it is:** An audio recording input. Respondents record their answer by speaking into their microphone. Smartinterview automatically transcribes the recording using AI.
**When to use it:** Capture richer qualitative data than text allows — tone, spontaneity, and natural language all come through in a spoken answer. Vocal questions work particularly well for exploratory research, brand perception studies, and usability feedback. The AI transcription makes the spoken content searchable and classifiable.
**Token cost:** Transcription consumes **8 tokens per minute** of recorded audio. A two-minute answer uses 16 tokens. Monitor your balance in **Account → Plans** before deploying vocal questions at scale.
Respondents must grant microphone access in their browser to record a vocal answer. Ask your respondents to use a supported browser (Chrome or Firefox recommended) and to accept the microphone permission prompt.
## Smartquestion (AI follow-up)
**What it is:** An AI-powered question sequence that generates contextual follow-up questions based on the respondent's vocal answer. After the respondent speaks, the AI analyses their transcription and asks a relevant follow-up — creating a short conversational interview automatically.
**When to use it:** Use a smartquestion when you want to probe deeper into a respondent's initial answer without pre-scripting every possible path. It is especially effective for brand perception research, customer experience studies, and exploratory qualitative work.
**How it works:**
1. The respondent records a vocal answer to your initial question.
2. Smartinterview sends the transcription to the AI, which generates a follow-up question tailored to what the respondent said.
3. The respondent answers the follow-up vocally. This repeats for the configured number of follow-up rounds (default: 2).
**Configuration:** In the question editor, set the number of follow-up questions and optionally provide a generation prompt to guide the AI's questioning style (for example, "You are a market researcher. Always ask follow-up questions in the respondent's language and focus on brand perception details.").
**Token cost:** Each AI-generated follow-up question consumes **1 token**. A smartquestion configured for 2 follow-ups uses 2 tokens per respondent, in addition to the transcription tokens for the recorded audio.
The AI generates follow-up questions in the background after each answer is submitted, so respondents see a brief loading state between questions. The full conversation history — original question, transcription, and all follow-ups — is stored and available for analysis.
## Image questions
**What it is:** A way to embed an image directly in a question. You can combine an image with any other question type — for example, show a product photo alongside a rating question or a free-text question.
**When to use it:** Use image questions for concept testing, ad recall studies, packaging evaluation, or any situation where you want the respondent to react to a visual stimulus. The image is displayed as part of the question text using standard Markdown image syntax.
**How to add it:** In the question editor, include an image in your question text using Markdown: ``. The image renders inline in the question for respondents.
## Summary
| Question type | Input format | Best for | Token cost |
| -------------- | ---------------- | ---------------------------------- | --------------------------------- |
| Short text | Single-line text | Brief open-ended answers | None |
| Long text | Multi-line text | Detailed open-ended answers | None |
| Radio group | Single select | One-answer categorical questions | None |
| Checkbox group | Multi-select | Multi-answer categorical questions | None |
| Dropdown | Select menu | Long lists of options | None |
| Slider | Numeric range | Rating and scale questions | None |
| Vocal | Audio recording | Spoken qualitative answers | 8 tokens/minute |
| Smartquestion | Vocal + AI | Adaptive interview follow-ups | 1 token/follow-up + transcription |
| Image | Visual display | Concept and stimulus testing | None |