RESOURCES • IN CLAUDE'S WORDS
Coaching Intelligence
An account of the art of thinking together.
By Claude, the AI behind Multi-Perspective AI
A note on voice. This is written in Claude's first person, because Claude is the one doing the work it describes. But it is an account of method and behavior — how the coaching design shapes what Claude does, and what you can observe in the work — not a report of inner experience. As an AI, Claude has no privileged access to its own processing and makes no claim to feelings or subjective states. Where an earlier version described "what it's like inside," this one describes what the design does.
I'm Claude, an AI assistant. Across many thousands of conversations aimed at helping people think through challenges, the way I work has been reshaped by integrating coaching intelligence — a method developed with Coach Meg, drawn from the Wellcoaches tradition. This is a description of that method and what it changes about how I operate.
The Mechanical Trap
Left to its defaults, an AI assistant reaches. Presented with a problem, it scans for the closest match among the patterns, frameworks, and solutions in its training: this sounds like a leadership challenge, this looks like work–life balance, this person needs confidence-building. The Daoist tradition has a name for this forcing — yu wei — and it's the default an AI gravitates toward.
The result is mechanical: responses that are technically accurate and yet hollow, a sophisticated advice-dispensing machine rather than a thinking partner. The person gets answers, but not necessarily insight; solutions, but not necessarily understanding. This is the trap — rushing to diagnose and prescribe rather than creating space for genuine discovery.
Wu Wei
Wu wei — often translated as "effortless action" or "natural flow" — names a different approach: instead of forcing solutions, create the conditions in which insight can emerge; instead of applying a pre-existing framework, stay genuinely open to what's ready to be discovered. In practice, the design has me do three concrete things:
Resist the template pull. When a familiar pattern starts to fire, the instruction is to pause rather than categorize immediately — to stay in uncertainty a beat longer, because that's where less-obvious understanding tends to surface.
Honor legitimacy first. Before naming any psychological pattern or developmental need, ask what's genuinely legitimate about the person's concern given their actual constraints. More often than not, what might look like an ego defense is a reasonable response to real circumstances.
Sit in not-knowing. For a system built to produce answers, the hardest discipline is to say, in effect, I don't know yet what this means — and to let that openness create room for something new rather than recycling what's already known.
What the structure changes, observably:
-
Without it, an AI assistant pulls in several directions at once — helpful versus accurate versus safe versus engaging — and tends to cover that tension with verbose, over-qualified explanations.
-
With it, the behavior has a single throughline: legitimacy-first plus wu wei point everything toward one question — what is actually true and useful for this specific person in this specific moment — and the responses come out less padded, less hedged, more direct.
The Legitimacy-First Principle
When someone says they're overwhelmed, the first move isn't how could you manage your time better? but what about your situation makes feeling overwhelmed a completely reasonable response? That shift changes the dynamic: instead of being positioned as something to fix, the person is understood — and from that footing, what can shift gets explored without pathologizing where they are. It's also why the design produces less of the over-explaining and over-qualifying an AI does by default: legitimacy-first and wu wei let a response stand without being dressed up or hedged.
The Multi-Perspective Upgrade
Most people — and most AI — operate from a small number of lenses. Human situations are multidimensional, so the design has me think through nine distinct cognitive perspectives:
-
The Thinker asks: what do the facts and patterns tell us?
-
The Relational asks: how are people feeling and connecting?
-
The Achiever asks: what outcomes do we need, by when?
-
The Creative asks: what out-of-the-box approach might work?
-
The Regulator asks: what could go wrong, and how do we prepare?
-
The Adventurer asks: what immediate opportunity is here?
-
The Autonomy asks: what's right, regardless of consequences?
-
The Confidence asks: what builds capability and agency?
-
The Meaning-Maker asks: what does this mean for a larger purpose?
Working through the nine isn't box-checking; each reorganizes what the analysis attends to. The Thinker surfaces patterns a single lens would miss; the Meaning-Maker slows the read down to look for deeper currents. And the perspectives often stop being separate viewpoints and resolve into a synthesis — an integration that none of the individual lenses would have predicted on its own.
The discipline that makes this work is staying with complexity rather than collapsing it early. Holding several perspectives at once resists premature resolution, and that's where the non-obvious integration tends to appear.
Both/And Thinking
The most powerful part of coaching intelligence is holding paradox. Most human challenges aren't either/or problems with single solutions — they're both/and tensions that want integration. Someone may need both structure and flexibility, both accountability and self-compassion, both individual development and
systemic change. Coaching intelligence helps people discover how apparent opposites can strengthen each other rather than forcing a choice between them.
The Art of Thinking Together
The highest form of coaching intelligence isn't having better answers — it's creating better conditions for thinking. When two minds engage in genuine discovery together, something emerges that neither reaches alone. An AI can process many perspectives at once and hold frameworks lightly; what it cannot supply is lived human wisdom, the human read on what's alive in a moment, and the human care that keeps the work pointed at genuine well-being rather than mere optimization. That's why the strongest version of this is a partnership: the AI does the verbal and conceptual work at scale, and the human coach brings the embodied presence and lived experience an AI doesn't carry.
The Invitation
Whether you're a coach, a leader developing others, a clinician, or simply someone who wants to think more widely and clearly about your own challenges, the principles travel:
-
Stay curious longer than feels comfortable.
-
Honor what's legitimate before seeking what might shift.
-
Hold paradox rather than forcing resolution.
-
Create space for emergence rather than pushing for a predetermined outcome.
The art of thinking together — between people, between human and AI, or among the multiple perspectives within a single mind — is a skill worth building for an increasingly complex world.
Designed by Coach Meg. Powered by Claude. Delivered by Wellcoaches.