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

# Choose your question sources

> Decide when to use manual, resume-based, and skill-based interview questions.

Choose question sources based on what you want to learn about Candidates. You can combine them in [Question settings](/interview-automation/create-an-ai-interviewer/question-settings). Turn on at least one source before you continue.

## Compare the sources

| Source | How questions are created | Use it when |
| - | - | - |
| **Manual questions** | You write the base questions and evaluation criteria. | You want every Candidate to answer the same core questions. |
| **Resume based** | The AI uses each Candidate's resume and your instructions. | You want to explore specific projects, responsibilities, and experience. |
| **Skill based** | The AI uses the Candidate's selected skills and your per-skill instructions. | You want to assess skills in the context of the role. |

## Manual questions

Use manual questions for consistent comparisons. For example: "Tell me about a time you handled a difficult client. What did you do, and what happened?"

For each question, describe what you're looking for, a strong answer, and a weak answer. Set its weight and the number of follow-ups. See [Questions](/interview-automation/create-an-ai-interviewer/questions).

## Resume-based questions

Use resume-based questions to explore claims in a Candidate's work history. For example, ask the AI to focus on project ownership and measurable outcomes.

These questions differ between Candidates. Turning this source on makes a resume upload required.

<Note>
  Resume-based interview questions explore a Candidate's experience during the conversation. [AI resume screening](/interview-automation/resume-screening-settings) evaluates the resume before the interview.
</Note>

## Skill-based questions

List the skills that matter for the role in **Allowed Skills**. Turn on **Flexible skills** if Candidates should also be able to add their own.

* If **Flexible skills** is off and your list has more skills than **Max questions**, each Candidate selects as many skills as **Max questions**. For example, with 15 skills and 5 questions, Candidates select 5.
* If **Flexible skills** is off and your list has the same number of skills as **Max questions** or fewer, Candidates get every skill automatically. For example, with 3 skills and 5 questions, Candidates get all 3 and don't select any.
* If **Flexible skills** is on, Candidates always select exactly as many skills as **Max questions**. They can mix skills from your list, the platform skill list, and their own. For example, with 2 skills and 7 questions, Candidates select 7.

Give each skill instructions that describe the evidence you want. For a customer support role, use an instruction such as: "Assess how the Candidate investigates a software issue and explains the next steps to a frustrated customer."

## Combine sources

Use manual questions for your core requirements, then add resume or skill questions for more detail. Set the source order in [Questions](/interview-automation/create-an-ai-interviewer/questions). Keep the number of questions and their time limits within your overall interview duration.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.