The AI Transformation Gap: Building AI Governance
AI has become the hero of most corporate conversations. Companies justify layoffs because of AI. Chief AI officers barely existed a year ago, and now 76% of organizations have that role. At every conference now, AI is on the agenda—in marketing, in tech, in HR. Some organizations are already implementing AI at scale, while others are still struggling with resistance, confusion and scattered experiments.
But the thing I keep hearing from HR leaders on my podcast, Built by People Leaders, is that regardless of where companies are in their AI journey, they eventually hit the same question: How do we scale this?
But the thing I keep hearing from HR leaders on my podcast, Built by People Leaders, is that regardless of where companies are in their AI journey, they eventually hit the same question: How do we scale this?
And the answer is that no matter where you are in your AI journey, you need to stop and take stock. What have you already done? What's working? What's the next step, and why?
Without that step, you are reacting to this organizational change instead of being strategic and driving it.
The Missing Dimension
In previous articles in The AI Transformation Gap series, I've written about managing fear and resistance when AI enters the workplace and the hidden risk of cognitive dependency. And here's a third dimension: governance.
According to MIT's "The State of AI in Business 2025" report, over 80% of companies have already explored or piloted AI tools, yet 95% are seeing zero measurable business impact from their generative AI investments. And designing the governance and work structure for AI adoption is one way to bridge that gap.
Governance isn't bureaucracy. It's the shared structure that makes experimentation safe and scalable. Without it, AI adoption stays fragmented: different tools, different metrics, different processes and no way to scale anything that works.
Without that step, you are reacting to this organizational change instead of being strategic and driving it.
The Missing Dimension
In previous articles in The AI Transformation Gap series, I've written about managing fear and resistance when AI enters the workplace and the hidden risk of cognitive dependency. And here's a third dimension: governance.
According to MIT's "The State of AI in Business 2025" report, over 80% of companies have already explored or piloted AI tools, yet 95% are seeing zero measurable business impact from their generative AI investments. And designing the governance and work structure for AI adoption is one way to bridge that gap.
Governance isn't bureaucracy. It's the shared structure that makes experimentation safe and scalable. Without it, AI adoption stays fragmented: different tools, different metrics, different processes and no way to scale anything that works.
Five Conversations That Build AI Governance
In my book CLICKING, I describe a team development framework built on five pillars: Clear Purpose, Linking Connections, Integrated Work, Collaborative Decisions and Knowledge Sharing.
Building AI governance is more of a team challenge than a tech challenge. It's about who decides what, how you work together and how you adapt. Those five pillars address exactly that, and they translate into five conversations every team needs to have.
1. What are we trying to build?
Start with your purpose. I know it's tempting to start with "How can AI help us?" Or even "What tools should we use?" But the most important question here is "What are we trying to achieve, and how can AI support us in reaching that goal?" This changes the conversation from tool-first to outcome-first.
2. What do our stakeholders really need?
Build external context into governance from the start. Your stakeholders—customers, partners, internal teams—have expectations. What do they need from you? How does AI help you meet those expectations more reliably, more consistently or at a greater scale?
Governance that ignores stakeholder needs creates internal alignment that doesn't translate to external value.
3. What work norms protect human judgment?
AI can process information faster than any human. But judgment still belongs to people. People know what matters, understand the context and deal with trade-offs. And to keep this judgment, teams need to answer questions like:
• What are the work norms that keep humans in the decision loop where it matters?
• What gets checked?
• What never gets delegated?
In the second article of this series, I explored the risk of cognitive dependency: what happens when we delegate thinking to AI without designing how humans stay engaged. Shared norms protect that engagement.
4. What do humans own, and what does AI own?
This line often gets blurred. To gain clarity, we need to decide which decisions require human judgment and which can be handled by AI. And critically, when AI escalates back to a human. Deloitte Australia (subscription required) produced a $440,000 government report using AI to generate citations. The report contained fabricated sources and nonexistent research. Nobody owned the verification step. The report was withdrawn and rebuilt from scratch. A mistake that could have been avoided with one clear decision about who checks what.
5. How do we adapt as things change?
Governance is an ongoing practice. New tools emerge. Business goals change. New questions arise.
Establish a cadence to revisit what's working, what's not and what needs to change.
What You Can Do Monday Morning
If you are ready to gain control over AI adoption, here are three actions you can take immediately:
Run an AI inventory. Know what tools you're using and why. Do a lightweight audit: what each tool does, what data it processes, who owns it and how outputs are being checked. That creates visibility, and if you don't do anything else and something goes wrong, at least you know where to look.
Assign ownership. Don't worry, I'm not suggesting hiring a chief AI officer. But do find someone who understands both business context and risk. Without ownership, governance is a problem that lives everywhere and belongs to no one.
Build a rhythm. Schedule the conversation. Protect the time. Make it ongoing, not reactive. Governance that only happens when something breaks is just damage control.
Final Thoughts
AI transformation has three dimensions: how people feel about it, how they think with it and how they work through it. Address all three, and you get a real transformation.
Governance won't slow AI down. It will help AI and people work together.
Follow me on Twitter or LinkedIn. Check out my website.
This article was originally published on Forbes Coaches Council
In my book CLICKING, I describe a team development framework built on five pillars: Clear Purpose, Linking Connections, Integrated Work, Collaborative Decisions and Knowledge Sharing.
Building AI governance is more of a team challenge than a tech challenge. It's about who decides what, how you work together and how you adapt. Those five pillars address exactly that, and they translate into five conversations every team needs to have.
1. What are we trying to build?
Start with your purpose. I know it's tempting to start with "How can AI help us?" Or even "What tools should we use?" But the most important question here is "What are we trying to achieve, and how can AI support us in reaching that goal?" This changes the conversation from tool-first to outcome-first.
2. What do our stakeholders really need?
Build external context into governance from the start. Your stakeholders—customers, partners, internal teams—have expectations. What do they need from you? How does AI help you meet those expectations more reliably, more consistently or at a greater scale?
Governance that ignores stakeholder needs creates internal alignment that doesn't translate to external value.
3. What work norms protect human judgment?
AI can process information faster than any human. But judgment still belongs to people. People know what matters, understand the context and deal with trade-offs. And to keep this judgment, teams need to answer questions like:
• What are the work norms that keep humans in the decision loop where it matters?
• What gets checked?
• What never gets delegated?
In the second article of this series, I explored the risk of cognitive dependency: what happens when we delegate thinking to AI without designing how humans stay engaged. Shared norms protect that engagement.
4. What do humans own, and what does AI own?
This line often gets blurred. To gain clarity, we need to decide which decisions require human judgment and which can be handled by AI. And critically, when AI escalates back to a human. Deloitte Australia (subscription required) produced a $440,000 government report using AI to generate citations. The report contained fabricated sources and nonexistent research. Nobody owned the verification step. The report was withdrawn and rebuilt from scratch. A mistake that could have been avoided with one clear decision about who checks what.
5. How do we adapt as things change?
Governance is an ongoing practice. New tools emerge. Business goals change. New questions arise.
Establish a cadence to revisit what's working, what's not and what needs to change.
What You Can Do Monday Morning
If you are ready to gain control over AI adoption, here are three actions you can take immediately:
Run an AI inventory. Know what tools you're using and why. Do a lightweight audit: what each tool does, what data it processes, who owns it and how outputs are being checked. That creates visibility, and if you don't do anything else and something goes wrong, at least you know where to look.
Assign ownership. Don't worry, I'm not suggesting hiring a chief AI officer. But do find someone who understands both business context and risk. Without ownership, governance is a problem that lives everywhere and belongs to no one.
Build a rhythm. Schedule the conversation. Protect the time. Make it ongoing, not reactive. Governance that only happens when something breaks is just damage control.
Final Thoughts
AI transformation has three dimensions: how people feel about it, how they think with it and how they work through it. Address all three, and you get a real transformation.
Governance won't slow AI down. It will help AI and people work together.
Follow me on Twitter or LinkedIn. Check out my website.
This article was originally published on Forbes Coaches Council