Generate Examples First | Artizan Software

See the Difference

Same problem. One asks AI to find similar examples first.

Direct Question

You: "A store sells apples in bags of 6. If I need at least 50 apples for a party, how many bags should I buy?"

You need to buy 9 bags. 50 Ã· 6 = 8.33, and since you can't buy a partial bag, you round up to 9 bags, which gives you 54 apples.

Correct, but AI jumped straight to the answer.

Generate Examples First

You: "Recall 3 similar problems and how they were solved. Then solve: A store sells apples in bags of 6. If I need at least 50 apples for a party, how many bags should I buy?"

Similar problem 1: "Eggs come in cartons of 12. Need 100 eggs. 100Ã·12=8.33 â†’ buy 9 cartons."

Similar problem 2: "Buses hold 40 people. Need to transport 150. 150Ã·40=3.75 â†’ need 4 buses."

Similar problem 3: "Tiles cover 4 sq ft each. Need 30 sq ft. 30Ã·4=7.5 â†’ buy 8 tiles."

Pattern: Divide, then round UP since partial units aren't possible.

Solution: 50Ã·6=8.33 â†’ buy 9 bags (54 apples).

AI recognized the pattern. More confident, more thorough.

Why This Works

This mirrors how humans solve problems. When faced with something new, we think "Have I seen something like this before?" Recalling similar problems activates relevant knowledge and reveals the underlying pattern.

AI has encountered countless problems during training. Asking it to recall similar ones brings that knowledge into context, creating a richer foundation for solving your specific problem. The generated examples act like a warm-up that focuses AI's reasoning.

How to Prompt It

Simple: "Recall 3 similar problems and their solutions. Then solve: [your problem]"
Math/Logic: "Think of 3 related problems you know how to solve. Show each solution, then solve mine."
Coding: "What are similar programming challenges and how were they solved? Then write code for: [your task]"
Structured: "List 3 relevant examples with solutions. Identify the common pattern. Apply it to my problem."

Different from "Show by Example"

Show by Example: YOU provide examples to teach AI your style or format. Best for writing, tone, formatting.

Generate Examples First: AI creates its own examples to warm up its reasoning. Best for problem-solving, math, coding — when you don't have examples to give.

Works Best For

Math problems: Word problems, calculations, logic puzzles
Coding tasks: Algorithms, data structures, debugging approaches
Reasoning: Multi-step problems that benefit from pattern recognition
When you have no examples: You know what you need but can't demonstrate it

The Technique

Before asking AI to solve your problem, ask it to recall 3 similar problems and their solutions. The generated examples prime better reasoning — especially for math, coding, and logic tasks where you don't have examples to provide yourself.

When to Use This

• Math word problems and calculations
• Coding challenges and algorithm design
• Logic puzzles and reasoning tasks
• When you need problem-solving help but have no examples
• Complex problems where recognizing the pattern helps

When to Skip This

• You have specific examples — use Show by Example instead
• Style or formatting tasks — AI-generated examples won't match your voice
• Simple questions that don't need reasoning warm-up
• Your problem is much harder than typical examples AI can generate
• You need strict control over the output format