The AI Transformation Gap: Preventing Cognitive Dependency
Sarah, my former client and the head of customer success at a SaaS company, was proud of her team. They were early adopters—enthusiastic, curious, willing to experiment. When AI tools became available, they didn't wait for a company-wide mandate. They jumped in.
Within months, they had transformed their workflow. Conversations were transcribed automatically. AI-generated insights and uploaded them into the CRM. Backlog items for product teams were created without manual input. Client call agendas, reporting, data analysis—all handled. The team finally had time. They weren't behind anymore. They had space for the strategic work they'd always wanted to do but never had capacity for.
Within months, they had transformed their workflow. Conversations were transcribed automatically. AI-generated insights and uploaded them into the CRM. Backlog items for product teams were created without manual input. Client call agendas, reporting, data analysis—all handled. The team finally had time. They weren't behind anymore. They had space for the strategic work they'd always wanted to do but never had capacity for.
They felt good—for a while.
Then something shifted. In team meetings, when discussing customers, the managers who used to know their accounts inside out couldn't recall details from recent conversations. They didn't know what was bothering their clients the most. They had to check the CRM for information that used to live in their heads. They struggled to prioritize backlog items because they'd lost their feel for what actually mattered to customers. Motivation dropped. Engagement dropped.
"We are just operators of AI," Sarah told me. They had lost meaning in their work.
The Cost Of Cognitive Dependency
In the first article of my AI Transformation Gap series, I explored how fear and resistance block AI transformation before it starts. But there is a second risk—one that emerges not when AI adoption fails but when it succeeds.
AI transformation has three dimensions that must be addressed simultaneously: how people feel about AI, how it affects the way they think and how teams work with it structurally. The second dimension is the hardest to recognize. It might feel like progress right up until performance starts to slip.
MIT Media Lab research shows that when people use AI first and then try to engage cognitively, brain connectivity is significantly weaker than when they think independently first and then bring AI in. The cadence matters—and most teams have it backward.
Then something shifted. In team meetings, when discussing customers, the managers who used to know their accounts inside out couldn't recall details from recent conversations. They didn't know what was bothering their clients the most. They had to check the CRM for information that used to live in their heads. They struggled to prioritize backlog items because they'd lost their feel for what actually mattered to customers. Motivation dropped. Engagement dropped.
"We are just operators of AI," Sarah told me. They had lost meaning in their work.
The Cost Of Cognitive Dependency
In the first article of my AI Transformation Gap series, I explored how fear and resistance block AI transformation before it starts. But there is a second risk—one that emerges not when AI adoption fails but when it succeeds.
AI transformation has three dimensions that must be addressed simultaneously: how people feel about AI, how it affects the way they think and how teams work with it structurally. The second dimension is the hardest to recognize. It might feel like progress right up until performance starts to slip.
MIT Media Lab research shows that when people use AI first and then try to engage cognitively, brain connectivity is significantly weaker than when they think independently first and then bring AI in. The cadence matters—and most teams have it backward.
But the cognitive impact doesn't stop at the individual level. Research from Northeastern University on AI's social forcefield found that AI reshapes how teams think together. When teams regularly work with AI, they start adopting AI language, AI framing and AI ways of structuring problems without realizing it. This influence persists even when AI is no longer in the room. Over time, it can erode the diversity of thinking that makes teams effective.
The warning signs are easy to miss. A team member who used to speak confidently about an account now hesitates. Priorities feel harder to set. Work feels more mechanical. These don't look like AI problems. They look like engagement problems, motivation problems, burnout. By the time the connection is made, the impact on performance is already real.
Changing The Sequence
Sarah's team didn't remove the AI. They changed when it entered the process.
Previously, the sequence was: Conversation happens; AI transcribes and generates insights; insights enter the CRM; team reviews what AI produced. The human was at the end of the chain, consuming output.
They reversed it. After each client conversation, team members first took time to reflect on their own—what stood out, what the client was really saying, what mattered most. Then AI generated summaries using both the human reflections and the transcript. The team reviewed and refined the AI suggestions.
The brain stayed in the loop. Ownership returned. And so did performance.
This single change—human thinking before AI input, not after—is the difference between using AI as a shortcut and using AI as an amplifier.
Three Practices For Keeping Your Team's Thinking Sharp
1. Think first, then prompt.
Before opening any AI tool, reflect on the problem, the client or the decision. Form your own view. Then bring it to AI. This practice is what keeps judgment and performance intact.
2. Challenge AI outputs as a team.
The Northeastern research shows AI homogenizes team thinking in ways nobody notices until decision quality drops. Build a habit of questioning AI outputs together: Does this match what we actually know? What is AI missing that we can see? What would we conclude without this? The goal is to keep your team's independent judgment alive while still using AI as a partner.
3. Define what you will protect.
In my book, Clicking, I describe an exercise called Keep-It-Up and Cut-It-Out behaviors—the practices teams want more of and the ones they want to eliminate. Applied to AI, this becomes a team conversation about what good human-AI collaboration looks like for your work: what you will always do as humans first, what you will never delegate to AI and what you will regularly challenge together.
You don't need strict AI policies to avoid cognitive dependency. What you need is honest conversations about what your team wants to protect and to build norms around it.
Using AI As An Amplifier
Sarah's team didn't become less efficient after they changed the sequence. They became more effective—because the people doing the work were still genuinely connected to it.
AI works best when it amplifies human thinking. The question every leader needs to ask is not how much your team is using AI. It's whether your team is still thinking—and whether that thinking is still genuinely theirs.
Follow me on Twitter or LinkedIn. Check out my website.
This article was originally published on Forbes Coaches Council
The warning signs are easy to miss. A team member who used to speak confidently about an account now hesitates. Priorities feel harder to set. Work feels more mechanical. These don't look like AI problems. They look like engagement problems, motivation problems, burnout. By the time the connection is made, the impact on performance is already real.
Changing The Sequence
Sarah's team didn't remove the AI. They changed when it entered the process.
Previously, the sequence was: Conversation happens; AI transcribes and generates insights; insights enter the CRM; team reviews what AI produced. The human was at the end of the chain, consuming output.
They reversed it. After each client conversation, team members first took time to reflect on their own—what stood out, what the client was really saying, what mattered most. Then AI generated summaries using both the human reflections and the transcript. The team reviewed and refined the AI suggestions.
The brain stayed in the loop. Ownership returned. And so did performance.
This single change—human thinking before AI input, not after—is the difference between using AI as a shortcut and using AI as an amplifier.
Three Practices For Keeping Your Team's Thinking Sharp
1. Think first, then prompt.
Before opening any AI tool, reflect on the problem, the client or the decision. Form your own view. Then bring it to AI. This practice is what keeps judgment and performance intact.
2. Challenge AI outputs as a team.
The Northeastern research shows AI homogenizes team thinking in ways nobody notices until decision quality drops. Build a habit of questioning AI outputs together: Does this match what we actually know? What is AI missing that we can see? What would we conclude without this? The goal is to keep your team's independent judgment alive while still using AI as a partner.
3. Define what you will protect.
In my book, Clicking, I describe an exercise called Keep-It-Up and Cut-It-Out behaviors—the practices teams want more of and the ones they want to eliminate. Applied to AI, this becomes a team conversation about what good human-AI collaboration looks like for your work: what you will always do as humans first, what you will never delegate to AI and what you will regularly challenge together.
You don't need strict AI policies to avoid cognitive dependency. What you need is honest conversations about what your team wants to protect and to build norms around it.
Using AI As An Amplifier
Sarah's team didn't become less efficient after they changed the sequence. They became more effective—because the people doing the work were still genuinely connected to it.
AI works best when it amplifies human thinking. The question every leader needs to ask is not how much your team is using AI. It's whether your team is still thinking—and whether that thinking is still genuinely theirs.
Follow me on Twitter or LinkedIn. Check out my website.
This article was originally published on Forbes Coaches Council