Three Questions to Think About When You Use AI

Directive thinkers ask these questions. Most people skip straight to the prompt.

1. "Am I trying to decide or produce?"

AI can't make decisions. It can produce variations.

❌ Don't ask: "Should we redesign our checkout?" (Requires judgment, user insight, business context)

✅ Ask: "Here are three checkout approaches. Generate variations of each." (Production. AI excels here.)

The rule: Decisions require your judgment. Production can use AI's speed.

2. "What does AI know vs. what do I know?"

AI knows: Patterns from training data. What usually happens.

You know: Your users. Your constraints. What hasn't been done. What matters.

AI suggests common solutions. You identify uncommon opportunities.

The rule: AI shows you the probable. You choose the valuable.

3. "When do I override probability?"

AI optimizes for "most likely." Interesting work requires "less probable but more valuable."

Example:

• AI suggests: The standard pattern everyone uses
• You decide: "We're going this direction instead because..."

The rule: AI gives you the safe bet. You make the interesting call.

The Pattern

❌ Weak: AI thinks → you execute
✅ Strong: You think → AI executes

❌ Weak: "AI, what should we do?"
✅ Strong: "Here's what we're doing. AI, help me do it faster."

You're the director. AI is crew.

Directors don't ask the camera operator what movie to make.

Same with AI.