Three Questions Test and Threshold Screen

What this is

The book defines the Transformation Threshold as the moment coaching crosses from "What do I do differently?" to "Who do I need to become?" and gives a three-question test for deciding whether something is AI-appropriate or human-required. This page reproduces the definition, the three domains it rests on, and the test as the book states them. It is a judgment aid. The book supplies no score.

Use when / Do not use when

Use when you are unsure whether something is AI-appropriate or human-required. The book says to use the questions every time you are deciding whether to engage AI or escalate. Do not use it to produce a number or an automatic routing result. The book gives no scoring, no cut-off and no count rule for this test.

The Instrument

The three domains the threshold sits between (book)

The book's reasons AI is weak at Transform: no lived experience, no genuine relationship, no stake in your outcome, and it cannot truly witness your struggle. The book's reasons AI excels at Analyze and Coach: unlimited processing capacity, no fatigue, no recency bias, infinite patience, always available, consistent quality.

The Transformation Threshold defined (verbatim)

"The threshold is the moment coaching crosses from 'What do I do differently?' to 'Who do I need to become?'"

• Below the threshold: Skills, habits, behaviors, techniques. AI can help.
• Above the threshold: Identity, meaning, purpose, ethics. Humans required.

Two dangers named by the book:

• Crossing the line: "AI that pretends to do transformation work builds false intimacy and fails when it matters most. People feel handled, not helped. Trust erodes. The critical moment gets missed."
• Not using AI below the line: "Wasting human energy on work that could be automated. Burning out managers on analysis and skill practice. Leaving transformation work under-resourced because humans are exhausted from doing what machines could do."

The Three Questions Test (verbatim)

"When you're unsure whether something is AI-appropriate or human-required, ask:"

1. Is this about skill or identity?

• Skill: "How do I give better feedback?" → AI can help.
• Identity: "Am I the kind of leader who avoids hard conversations?" → Human needed.

2. Is this about pattern or meaning?

• Pattern: "What's causing my deals to stall at the proposal stage?" → AI can help.
• Meaning: "Why does this keep happening to me? What am I not seeing about myself?" → Human needed.

3. Is this about repetition or breakthrough?

• Repetition: "I need to practice this presentation ten times and get feedback each time." → AI can help.
• Breakthrough: "I need to fundamentally change how I show up in high-stakes situations." → Human needed.

"The questions are simple. The implications are profound. Use them every time you're deciding whether to engage AI or escalate to a human."

The book's illustrative pair (summary, book illustration, unverified, not for ROI or proof claims)

Same manager, same week. One conversation is a rep who defaults to pitching too early on discovery calls: a framework, a role-play and feedback; the book calls this skill work that AI could largely have done. The other is a senior rep questioning whether she wants to stay in sales although the numbers are good; the book calls this identity work AI cannot touch. Only the second required the manager specifically.

Related threshold wording elsewhere in the book (cross-reference)

When unsure whether the threshold has been crossed: "Is this about what to do, or about who to be?" / "Is this about execution, or about direction?" / "Is this about skill, or about something deeper?" "If it's the second option, you need a human." (Reproduced in "Escalation Decision Tree and Non-Delegable List".) The book also tells individuals to use the three questions test when they catch themselves asking AI identity questions.

How to run it (Performance Lab suggested practice)

1. State the question or situation in the person's own words.
2. Ask the three questions in order. For each, write which side it falls on and quote the words that show it.
3. If any answer is unclear, record it as unclear and ask the client more; do not guess.
4. Record what you decided to do (AI support, human conversation, or both), who decided and why. The book gives no rule for a mixed result, so this is a coach judgment.
5. Where a human is needed, go to "Escalation Decision Tree and Non-Delegable List" for escalation and "Human Domain Protections and Escalation Protocol" for what protected human time looks like.

Record sheet (Performance Lab suggested practice)

Situation (client's words): ______________________