TL;DR: Generic AI gives you a confident list of best practices. A Lean AI coach asks questions first, helps you understand your specific situation, and lets the advice follow from there. Same question, completely different orientation.
Ask ChatGPT a question about Lean in healthcare and you'll get an answer. Usually a confident one, organized into a tidy numbered list, with bold headers and an authoritative tone.
The problem isn't that the answer is wrong. Often, it's mostly right. The problem is that giving someone the answer is the opposite of what a good Lean coach does.
What Happens When You Ask Generic AI for Help
Try asking ChatGPT:
“How do I get my team to participate in daily huddles?”
You'll get something like a five-step plan: set clear expectations, keep it short, assign roles, use a visual board, follow up on action items. All reasonable. None of it wrong.
But nobody asked you what your huddle looks like today, or why people aren't participating. Is the problem the format? The timing? The psychological safety of the room? Whether anyone has ever seen something useful come out of it?
A generic AI treats your question as a request for information. A Lean coach treats it as the beginning of a conversation.
Why Coaching Matters More Than Answers
This isn't a philosophical preference. It's rooted in how improvement actually works.
If I hand you a best practice, you might try it. When it doesn't work — and it often won't, because your context is different from wherever it originated — you're stuck. You did what the expert said. It didn't help. Now what?
Anyone who has worked with a Toyota-trained coach knows the experience: you ask them what to do and they answer with a question. Not because they don't know, but because you remembering their answer is less valuable than you working through the problem yourself. A coach who helps you understand the problem deeply is helping you build capability. You can design your own experiment, and when it doesn't work perfectly the first time, you know enough to adjust.
If you're giving somebody answers, it robs them of their ability to develop their own problem solving skills. That's something I learned from John Shook a long time ago, and I've heard it from other Toyota folks. Toyota is concerned with developing people (a long-term view) not just solving the short term problem.
Related posts:
Long-Term Thinking: The Cornerstone of Lean and Principle #1 of the Toyota Way
Toyota as a People Development Company: Why Lean Starts With Developing People, Not Tools
When a Question Is Actually a Directive
What the Lean Hospitals AI Coach Does Differently
When you switch the Lean Hospitals AI Coach from Facts mode to Coach mode, it stops dispensing answers and starts acting as a thinking partner.
Ask it the same huddle question and it might respond:
“Tell me about your current huddle — how long does it run, who's there, and what happens with the issues that get raised? I want to understand what's working before we talk about what to change.”
It limits itself to two or three ideas instead of a long list. It suggests small experiments rather than prescribing solutions. It follows up with a question that moves the conversation forward.
Coaches ask questions not because they don't know an answer from their past experiences. It's because the answer without context is almost always the wrong answer for your situation.
Also, the generalized industry-agnostic “Lean Coach” that I also launched behaves the same way.
The Guardrails That Shape the Coaching
Generic AI has no perspective. It will agree with whatever frame you bring to the conversation. Ask it how to use Lean to cut headcount and it will help you do exactly that — cheerfully reinforcing one of the most damaging misconceptions in healthcare improvement.
Related post: “How Do I Use Lean to Reduce Headcount?” — Why ChatGPT's Answer Should Worry You
The Lean Hospitals AI Coach has philosophical guardrails — affirmative principles that shape every response, and common distortions it's designed to push back on. It knows that Lean is not a cost-cutting program. It knows that psychological safety isn't about comfort — it's about enabling people to speak up and challenge the status quo without fear. It knows that “implementing Lean” as a finite project misses the point.
In Coach mode, these guardrails work together with the Socratic approach. The AI doesn't ask questions for the sake of asking. It asks questions that steer you toward systems thinking, toward Respect for People, toward understanding root causes rather than jumping to solutions.
A Real Example: “My CNO Cancels Every Gemba Walk”
Here's where the difference becomes concrete.
Tell ChatGPT: “My CNO says she supports Lean but cancels every gemba walk and never asks about improvement work.” You'll get five numbered recommendations — clarify expectations, frame Lean in terms of organizational goals, make it easy for her to engage, build a case for improvement work, find other champions.
Notice the irony: the AI is telling you to “have a conversation” and “ask her directly” — while itself skipping that exact step. It gave you five recommendations without asking a single question about the situation.
Generic AI will advise you to be curious while being entirely incurious itself.
Ironic.
The Lean Hospitals AI Coach responds differently:
“When the CNO cancels, does she explain why, or is it just dropped from the calendar? And when you say she never asks about improvement work, is that just in your team, or do others notice the same thing?”
Two questions. No advice yet. And an explicit statement: “Understanding a bit more about the context will help me offer ideas that actually fit your situation.”
The answers to those two questions would lead to very different coaching conversations. If she's canceling because of competing crises, that's one problem. If gemba walks just quietly disappear from the calendar, that's a different one. If the whole organization notices, that's a systemic issue. If it's just your team, maybe the relationship needs attention first. ChatGPT's advice is the same regardless. The Coach's next move depends entirely on what you say.
ChatGPT seems to think you wanted an answer, and you'll be done. My AI leans into the start of a conversation.
Try It Yourself
If you've worked with a skilled Lean sensei, this approach will feel familiar. Good coaches don't walk into a gemba visit and start telling people what to fix. They ask questions. They observe. They help you see what you've been walking past.
That's the experience I wanted to build into the Lean Hospitals Coach. Not a search engine for Lean best practices — a thinking partner that helps you work through your own challenges, grounded in the principles from “Lean Hospitals” but applied to your context.
The best way to see the difference is to try the same question in both tools. Ask ChatGPT something about Lean in healthcare, then ask the Lean Hospitals AI Assistant the same thing in Coach mode. You'll notice the difference in the first response. Here are some prompts to get you started:
- “My boss wants me to create a Lean roadmap for the next two years.”
- “My staff say they don't have time for improvement work.”
- “We did a Lean training but nothing changed afterward.”
- “I'm a new Lean facilitator and my first kaizen event is next week. What should I do?”
- “Our ED wait times keep getting worse despite our improvement efforts.”
Try the full platform for free — for 48 hours — no account or credit card required. Try the industry-agnostic version.
Mark Graban is the author of Lean Hospitals and the creator of the Lean Hospitals AI Coach.
One takeaway I have from this post is that lean leadership is focused on asking questions to develop people’s understanding, rather than simply giving answers or how to do things. Instead of jumping to the answer, a lean coach develops people so they can fully understand and will be able to solve it on their own. This is important because when you are teaching someone, you want them to eventually be able to do it themselves and do it well. In my green belt project, this can be useful when asking deeper questions so that we can help teach the company and class so they can fully understand.