> ## 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.

# Questions

> Review question order and the prompts the AI will use.

Review the final question flow. This is the best place to confirm the Interviewer still sounds intentional after edits.

<Tip>
  **Quality check the interview flow:**

  * Put the highest-signal question sections first.
  * Tighten vague wording before you save so evaluation remains consistent.
  * Check generated counts if the interview now needs more or less depth.
</Tip>

### 1. Review dynamic questions

For sources like **Resume questions**, verify the settings:

* **Max questions**: The number of questions generated automatically for each shortlisted Candidate from their resume evidence.
* **Instructions**: Verify the specific prompt the AI uses to pull questions.

### 2. Add manual questions

The AI will ask these manual questions during the interview. For each manual question, you must complete the following evaluation criteria framework:

| Component             | Purpose                                                                                              | Details & Limits                              |
| --------------------- | ---------------------------------------------------------------------------------------------------- | --------------------------------------------- |
| **Question**          | The prompt the AI asks the Candidate.                                                                | Must be specific and clear.                   |
| **Looking for**       | The specific traits or evidence you want to hear.                                                    | Guides the AI's scoring engine.               |
| **Strong answer (5)** | Concrete examples of a top-tier response.                                                            | Sets the benchmark for a perfect score.       |
| **Weak answer (1)**   | Red flags or lack of specific evidence.                                                              | Sets the floor for a failing score.           |
| **Weight**            | Importance level (1 to 10).                                                                          | 1 is informational, 9 to 10 is role-critical. |
| **Max follow-ups**    | How many times the AI can dig deeper.                                                                | Must be between `0` and `2`.                  |
| **Time limit**        | You can set a time limit for each question, and Candidates will see this timer during the interview. | Between `0.5` (30 seconds) and `10` minutes.  |

### 3. Weight & follow-up guidance

* **Weight 1**: Informational only (excluded from overall score).
* **Weight 5 to 8**: Core competencies.
* **Weight 9 to 10**: Role-critical differentiators.
* **Max follow-up (0)**: Use for straightforward factual checks.
* **Max follow-up (1 to 2)**: Use for behavioral, competency, or dynamic resume/skills questions to ensure the AI gets enough evidence without derailing the time limit.

### 4. Question-writing formula

Use this structure to ensure the AI gets the best data:

1. Ask for a specific past situation.
2. Ask for actions taken.
3. Ask for a measurable outcome.

* **Better**: "Tell me about a time you improved conversion in a low-performing funnel. What did you change and what metric improved?"
* **Avoid**: "How do you think about conversion optimization?"

### 5. Interview prompt-aware best practices

The AI interviewer is programmed for specific behaviors. To align your questions with its logic:

* **Neutrality**: The AI stays neutral (no praise).
* **Evidence-based**: It will utilize its follow-up allowance (up to 2 times) to request specific evidence when answers are vague.
* **Scoped**: Keep each question focused on exactly one competency.