The Problem With AI Isn’t Tools. It’s Stinkin’ Thinkin’ in Leadership.
Most organizations are doing “AI work” right now—pilots, training, licenses, experimentation.
And still, the value isn’t compounding.
Here’s what I think is actually happening:
The problem with AI isn’t tools. It’s stinkin’ thinkin’ in leadership and problem solving.
“Stinkin’ thinkin’” is old recovery language for the mental loops that keep you stuck in yesterday’s patterns—even when you’re trying to change. And that’s exactly what I see in organizations: new technology layered onto old thinking.
We’re in a transition: AI-integrated vs AI-native
Right now most companies are in an AI-integrated phase: we bolt AI onto existing roles, processes, and approval chains.
An AI-native workplace is where we’re headed: the organization is designed assuming AI is part of the system—built into how work flows, how decisions get made, and how leaders lead.
The difference isn’t the model.
The difference is whether leaders keep asking the old questions… or learn to ask new ones.
The old questions create activity—not advantage
Old questions sound responsible:
- “Which tool should we choose?”
- “How do we get adoption?”
- “What’s our policy?”
- “How many use cases do we have?”
But those questions mostly produce rollout behavior.
AI-native leadership is a design discipline. It starts with different questions.
The AI-native question set (steal this)
If you want AI to become built-in instead of bolted-on, start here:
1) What decision are we trying to improve?
Not what task we can automate—what decision needs to get faster, cleaner, more consistent, less risky.
2) What does “good” actually mean for that decision?
Quality bar. Tradeoffs. Failure modes. What “wrong” looks like. This is where most teams stay vague—and AI punishes vagueness.
3) What must remain human-owned?
Judgment under uncertainty. Exceptions. Relationships. Ethics. Accountability. Name it and assign it—don’t leave it implied.
4) Where does accountability sit once AI touches the work?
Who signs off? Who escalates? What gets audited? What is explicitly not allowed?
5) What do we need to teach so people can work with AI without losing thinking?
Most orgs train on features. AI-native orgs train on how to think:
- problem framing
- critique and verification
- decision hygiene (inputs, assumptions, risk)
- escalation judgment (when to trust, when to verify, when to stop)
If you don’t teach that, you get the two predictable outcomes:
- over-trust (outsourcing thinking)
- under-trust (quiet refusal)
Both look like “adoption issues,” but they’re actually leadership issues.
Any maybe most importantly
6) Where are the areas of greatest impact that AI can add immediate value?
This will not necessarily be your biggest process problem to solve but your area of greatest impact to visible KPO’s life revenue, customer acquisition, etc….
Practical Move of the Week: Run an Impact-First Workflow Reset
Pick one workflow that touches revenue or margin this quarter (sales qualification, renewals, customer resolution, pricing approvals, forecasting, delivery throughput—something tied to real outcomes).
Then run this five-question reset with the team:
- What business outcome will this workflow move in the next 90 days?
- What decision inside this workflow is the constraint right now?
- Where can AI generate drafts / triage / recommendations to improve that decision?
- Where must human judgment remain mandatory (and why)?
- How will we measure impact weekly—without hiding behind “adoption”?
Rule: If you can’t answer #1 and #5 clearly, you’re not doing an AI project—you’re running a tool rollout.
To receive thought provoking-forward looking updates about technology integration, leadership and scalability, Click Here to Subscribe.
To take a look at last weeks’ post Click Here
Estimated reading time: 3 minutes