Login

Creating Research Outlines in Perspective AI

A research outline is a set of instructions that tells Perspective AI how to conduct an interview. It defines your research goals, target participants, interview flow, and the questions that need to be explored. Creating a well-structured outline ensures your AI-powered interviews deliver actionable insights aligned with your business needs.

Before You Begin

Before writing any questions, define what you need to learn by asking yourself:
  • What decision will this research inform? Identify the specific product, design, or business decision waiting on these insights.
  • What hypothesis or assumptions are we testing? Clarify what you believe to be true and what needs validation.
  • What would success look like for this interview? Define clear outcomes that would make the research valuable.
These questions help you craft a focused research question that generates a useful outline.

Creating Your First Research Outline

1. Enter a specific research question

Navigate to the research creation interface and enter a detailed research question. Be specific about who you want to interview and what you want to learn.
Example format:
I want to interview [target audience] about [specific topic] 
so that I can learn [desired outcome].
Sample question:
I want to interview existing customers about their experience using 
our beta AI feature so that I can learn what we need to improve 
before launching.

2. Click "Create Research"

Perspective AI will assess its confidence level based on your input. If the system needs clarification about your goals or target audience, it will ask follow-up questions.

3. Answer any follow-up questions

Respond to clarifying questions to help the system understand your research objectives. After answering a few quick questions, Perspective AI will generate your research outline.

Understanding Your Research Outline Components

Your generated outline includes several key elements:

Interview Type

Defines the research methodology (e.g., exploratory, evaluative, generative). See the interview types reference for detailed explanations of each type and their intended purposes.

Description

Clearly states what is being researched and by whom, providing context for the interview.

Research Question

The core question formulated to address the key business and user needs identified during planning.

Goals

The key objectives derived from your research question. These guide what the interview should accomplish.

Participants

A description of the target audience profile, including relevant demographic, behavioral, or firmographic criteria.

Interview Guidelines

Define the flow and tone instructions for the AI interviewer. Think of these as guardrails that shape how the interview agent conducts conversations.

Optional Components

Research outlines can also include:
  • Mandatory questions – specific questions that must be asked in every interview
  • Resources – reference materials for the interviewer
  • Form fields – structured data to collect from participants
See individual help articles for details on each component.

Refining Your Research Outline

Dictate changes to Perspective AI

After reviewing the initial outline, you can request modifications conversationally. Each change is evaluated and generates a new version of the outline.
Example changes:
  • "I want to interview enterprise customers specifically"
  • "Add a goal about understanding feature adoption barriers"
  • "Make the interview tone more conversational and less formal"

Test the interview yourself

Try the interview as if you were a participant. This hands-on experience often reveals:
  • Questions that need clarification
  • Flow issues in the conversation
  • Missing topics or goals
  • Tone adjustments needed

Iterate until satisfied

Repeat the cycle of dictating changes and testing until the interview meets your standards. Each iteration helps you fine-tune the experience for actual participants.

Next Steps After Creating Your Outline

  1. Review the complete outline – Ensure all components align with your research goals
  2. Test the interview – Experience it from the participant's perspective
  3. Make final adjustments – Dictate any last changes based on your test
  4. Invite participants – Once satisfied, begin recruiting and scheduling interviews

Best Practices

Start broad, then narrow. Your initial research question should capture the big picture. Use the iteration process to add specificity.
Test before launching. Always conduct a practice interview yourself. You'll catch issues that aren't obvious when reviewing the outline.
Focus on learning, not validation. Frame your research question around what you need to learn, not what you hope to confirm.
Iterate based on feedback. After your first few real interviews, review the conversations and refine the outline if needed.
Be specific about participants. The more detail you provide about your target audience (e.g., "enterprise customers" vs. "customers"), the better Perspective AI can tailor the interview.

Common Pitfalls & Fixes

Vague research questions → Add specific details about audience, context, and desired outcomes. Instead of "interview customers about the product," try "interview free-tier users about their upgrade decision process."
Too many goals → Focus on 3-5 key objectives. Too many goals can make interviews unfocused and exhausting for participants.
Overly rigid guidelines → Interview guidelines should be guardrails, not scripts. Allow the AI interviewer flexibility to have natural conversations.
Skipping the test interview → Always test yourself first. Reading an outline is different from experiencing the interview flow.