Module 16 · Tier 9 — Pioneer

Developing Your AI Thesis

How to develop a specific, defensible, original view of how AI reshapes your domain — and why this matters more than staying current.

20 min Tier 9 — Pioneer

Why This Matters

Every pioneer practitioner has a thesis — a specific claim about where things are going and why. Not a prediction (predictions are fragile) but a framework that explains what's happening, what it implies, and what to do about it. Without a thesis, you're reacting to events. With a thesis, events become evidence. Developing yours is the foundational intellectual work of Pioneer-tier practice.

An AI thesis is not a prediction ("AI will automate 40% of jobs by 2030"). It's not a position statement ("AI is transforming business"). It's a structured argument: here is what I believe about how AI is reshaping a specific aspect of my domain, here is why I believe it, here is what it implies, and here is what would change my mind.

The test of a thesis: if a thoughtful, informed person in your field read it, would they find something to disagree with? If not, it's not a thesis — it's a truism. A good thesis is specific enough to be wrong, which is what makes it valuable.

One to three sentences that state the distinctive insight. Not "AI is changing X" but "The specific mechanism by which AI is changing X is Y, and this is underappreciated because Z." Examples from practitioners: "AI doesn't replace professional judgment — it reveals which professional judgments were actually arbitrary." "The primary competitive effect of AI in our sector will be acceleration of the winner-takes-most dynamics already in motion, not the creation of new entrants."

Full access is for AIQ members

Unlock all 56 lessons, the certificate pathway, and the SociA|~ community.

56 lessons across all three course tracks
AIQ certification on completion
SociA|~ Society community access

Unlock full access →

The Concept

What a Thesis Is (and Isn't)

The Architecture of a Strong AI Thesis

The Core Claim

The Evidence Base

What leads you to believe this? Primary research you've consumed, operational experience, patterns you've observed, conversations with practitioners, first principles reasoning. A strong evidence base includes things you've personally observed that aren't in public discourse yet.

The Implication Stack

If your core claim is true, what follows? Work through the implications for your organisation, your sector, your own role, and the people you lead. A thesis without implication stack is interesting but not actionable.

The Update Conditions

What evidence would change your mind? Pioneers distinguish themselves from ideologues by being genuinely updateable. Stating your update conditions in advance is a discipline that keeps the thesis honest and earns intellectual credibility with serious practitioners.

How Theses Develop: The Iteration Cycle

A thesis is never finished — it's a living document. The cycle:

Draft: Write a rough version based on current thinking
Test: Share with 2-3 people who will seriously challenge it
Absorb: Note the strongest objections and where they land
Revise: Update the thesis, not to accommodate pushback but to become more precise in response to it
Publish: Share a version publicly — article, talk, post, conversation — and invite further challenge
Repeat: The public reaction generates the next round of revision

The output of this cycle is a thesis that has been stress-tested against serious thinking. This is qualitatively different from a view you've developed in private.

Three example thesis structures from different sectors

Professional services thesis: "AI will bifurcate the professional services market faster than firms expect. The middle tier — competent generalists who charge premium rates for work AI can replicate — will compress rapidly. But the demand for judgment-intensive work that AI cannot do will expand, and currently underserved. Firms that reposition upstream will grow; those that compete in the middle on efficiency will consolidate or exit."

Healthcare leadership thesis: "The primary near-term AI opportunity in clinical settings is not diagnostic accuracy — it's administrative load. Clinicians spend 35-40% of their time on documentation and coordination work that AI can largely absorb. The organisations that redirect that time to patient contact will see outcomes improvements and clinician retention effects that build compounding advantages."

Financial services thesis: "Regulatory bodies will move from rules-based to AI-assisted compliance oversight within 5 years. Firms that have built genuine AI governance capability will find this a competitive advantage; firms that have performed compliance without building underlying governance will face higher regulatory scrutiny, not lower."

Notice what each thesis has: a specific mechanism, a specific timeframe or condition, and a specific implication that varies by organisation type.

Hands-On Exercise

Write, test, and iterate your AI thesis

ClaudeAny AI assistantDomain-specific research sources

This is a multi-session exercise. Do not rush it. Session 1 (today, 20 minutes): 1. Write your core claim. One paragraph. What do you believe about how AI is reshaping your specific domain that differs from the mainstream view? 2. Write your evidence base. What have you seen, read, or experienced that leads you here? 3. Write your implication stack. If your claim is true, what are the top three implications for your organisation or sector? 4. Write your update conditions. What would change your mind? Between sessions: Share the draft with two people — one who knows your domain, one who will challenge the assumptions. Session 2 (after feedback): 1. What was the strongest objection? Was it about the claim, the evidence, or the implication? 2. Revise the thesis to address the objection — not by removing the claim but by becoming more precise. 3. Identify one venue where you could share a version of this publicly.

The hardest part is writing the update conditions. If you can't articulate what would change your mind, you don't have a thesis — you have a conviction. Convictions are less useful than theses.

Active Recall

Before moving on — close this lesson and answer these from memory. Then come back and check. Testing yourself (not re-reading) is how this sticks.

1 What distinguishes an AI thesis from a prediction or a position statement? What are the four structural components?

2 Why is public iteration essential to developing a robust thesis, and how does it improve quality differently than private revision?

Reflection

Do you currently have a thesis about AI in your domain — something specific enough that an informed peer could disagree with it? If not: what's the question you most want to have a view on, and what would it take to develop that view to the point of being worth sharing?

Key Takeaway

An AI thesis is a structured, revisable argument with four components: core claim, evidence base, implication stack, and update conditions. Its value comes not from being right but from being specific enough to be improved. Public iteration — sharing before it's finished, absorbing serious challenges, revising — is what transforms a private opinion into a durable intellectual contribution.