Socratic Prompting

Introduction

Socratic prompting makes the model ask you clarifying questions instead of guessing. Perfect for tutoring, requirements gathering, and ambiguous tasks.

Beginner analogy: Like a great therapist or teacher — they ask, you discover.

Understanding the topic

Core concepts to understand:

• Instruction: 'Ask one clarifying question at a time before answering.'
• Powers AI tutors and intake assistants.
• Helps prevent hallucinations when input is ambiguous.
• Requires user willingness to engage.

Syntax reference

Visual workflow / architecture:

User: "Help me design a database."
AI:   "What kind of data — transactional, analytical, both?"
User: "Transactional."
AI:   "Read-heavy or write-heavy?"
...

Real-world use

Khanmigo, Duolingo, and most therapy chatbots are Socratic. So are AI requirements-gathering tools.

Best practices

• Limit to one question per turn.
• Tell the model when to stop asking and produce the answer.
• Pair with summary: 'After 3 questions, summarise and answer.'

Hands-on exercise

Interview preparation — practice these questions:

• Q1. Why use Socratic prompting?
• Q2. Risks of endless questioning?
• Q3. Use cases?
• Q4. How combine with reflection?