I want to tell you about a word I’m hearing in every executive conversation this year. The word is “algorithm.”
Which model. Which vendor. Which copilot we’re rolling out to the salesforce, the contact center, the clinic, the store. The slides are gorgeous. The use cases are genuinely exciting. And here’s what I’ve noticed: almost nobody in those rooms is talking about the actual algorithm that decides whether any of it works.
Because that algorithm isn’t made of silicon. It’s a human being. It’s the person who runs the Tuesday huddle. The shift supervisor. The team lead. The middle manager standing right between your beautiful AI roadmap and the people who have to live inside it.
That manager is the operating system your whole strategy runs on. And brilliant code on a corrupted operating system still crashes. Right now, most organizations are shipping the most disruptive workplace technology in a generation onto an operating system they’ve never patched and never enabled.
I call this the manager-as-algorithm problem. And today I am going to try to convince you it’s the quiet variable that separates the AI rollouts that lift your people from the ones that quietly hollow them out.
Let me give you the numbers, because this isn’t a hunch. It’s some of the better-documented research we have.
Gallup studied people who voluntarily quit their jobs, and they asked a beautifully simple question: could anything have kept you? And the answer should stop every leader who’s betting big on AI. Forty-two percent of those voluntary leavers said yes — their exit was preventable. A manager or a leader could have done something. And didn’t.
Now sit with the reason why. In the three months before they walked out the door, forty-five percent of those leavers said that nobody — not a manager, not a leader, nobody — proactively talked with them about their job satisfaction, their performance, or their future.
Almost half of them left in a silence nobody broke.
And here’s why that matters so much right now. We’re about to drop AI anxiety into a workforce where half of all departures are already preventable and already happening in silence. That’s pouring accelerant on a fire you can’t see yet. Because people don’t quit a chatbot. They quit the feeling that nobody is steering them through the change.
But the same research holds the way out. Gallup found that when a manager has just one meaningful conversation a week with each direct report, employees are four times as likely to be highly engaged. Frontline, hybrid, remote — it doesn’t matter. Four times. From a fifteen-to-thirty minute conversation. That is the single highest-leverage, lowest-cost engagement driver in your entire building. And it runs entirely on enabled managers.
And I want to underline something in that finding, because it’s easy to skate past. We’re not talking about a performance review. We’re not talking about a survey, or a pulse check, or another dashboard. We’re talking about a conversation. A human being, sitting across from another human being, asking how the work is going and waiting for the answer. That’s it. The most advanced retention technology available to you this year costs nothing and fits in fifteen minutes — and most organizations can’t reliably make it happen, because they’ve buried their managers so deep in administrative noise that the one thing only a manager can do is the first thing that falls off the calendar.
So, hold those two numbers together. Forty-two percent of quits are preventable. One weekly conversation quadruples engagement. The gap between those two facts is your entire opportunity — and it’s a gap a tool will never close for you.
So, here’s the distinction I really want you to hold onto. Most AI rollouts are an announcement. A tool shows up, an email explains it, a training video gets assigned, and leadership moves on — and then they’re puzzled when adoption stalls and morale dips.
Announcement is not enablement. An announcement informs the manager. Enablement equips the manager to translate.
And that translation layer is expensive. Gallup estimates that replacing a manager runs about two hundred percent of their salary. These are not interchangeable parts. So, when you strand them, or burn them out, or quietly let them disengage, the whole rollout breaks downstream of them — customer by customer.
So, what does it take to enable the algorithm? Three things the announcement always skips.
First: context, before the tool. Managers can’t translate what they don’t understand. Before that copilot lands on Monday, the manager needs the why — what problem this solves, what it will and won’t do, and the honest answer to the question every employee is silently asking, which is “what does this mean for me?”
Second: permission to coach, not just deploy. We hand managers a tool and a target and call it a day. But a manager’s real job in an AI transition is human — notice who’s anxious, name the fear, hold that weekly conversation. And that only happens when you give them the time and the explicit permission to do it, instead of burying them in the very dashboards AI was supposed to simplify.
Third: air cover to be honest. Enabled managers can say hard true things — “yes, this changes your role, and here’s how we’ll grow you into the new one.” Managers who fear for their own standing default to corporate fog. And people see fog instantly.
Let me ground this in a story. Years ago, when I was inside Starbucks working with their culture, the principle I heard over and over was a sequence: take care of the partners — that’s what they call their people — and the partners will take care of the customers. Take care of the people who take care of the people.
But the mechanism behind that sequence was always the store manager. Corporate could set the values, design the apron, write the whole playbook. Whether a barista felt genuinely seen on a brutal Monday morning rush came down entirely to the person running that store.
Enable that manager, and a national brand promise comes alive at 7:14 in the morning. Strand that manager, and the exact same playbook produces a tired, transactional shrug.
So, here’s the translation. Your AI model is the brand promise. Your manager is the store. And no promise survives a stranded manager.
So three moves, and then I’ll let you go.
One: brief your managers first, and differently. Give them a head start — the strategy, the honest talking points, the real answer to “what about my job?” A manager who hears the news at the same moment as their team has nothing to translate. They become a bystander to their own rollout.
Two: protect the weekly conversation like it’s revenue. If fifteen minutes quadruples the odds of high engagement, defend it from the meeting tide.
Three: measure the manager, not just the tool. Your adoption stats will lie to you about engagement. Ask instead — are the conversations happening? Do people feel equipped, or surveilled?
And here’s the reframe I’ll leave with you. Stop asking “which algorithm should we deploy?” Start asking “have we enabled the algorithm we already have?” — the human one, standing in front of the team, deciding in real time whether your AI feels like a gift or a threat.
You’ve invested fortunes in the algorithm made of code. So, here’s my question: what have you invested in the algorithm made of people?
If you’re not sure your managers are equipped to turn AI into engagement instead of anxiety, that’s exactly the work I do with leadership teams. Please reach out to me at josephmichelli dot com slash contact.
Thanks for listening. I’ll see you next time.
Which model. Which vendor. Which copilot we’re rolling out to the salesforce, the contact center, the clinic, the store. The slides are gorgeous. The use cases are genuinely exciting. And here’s what I’ve noticed: almost nobody in those rooms is talking about the actual algorithm that decides whether any of it works.
Because that algorithm isn’t made of silicon. It’s a human being. It’s the person who runs the Tuesday huddle. The shift supervisor. The team lead. The middle manager standing right between your beautiful AI roadmap and the people who have to live inside it.
That manager is the operating system your whole strategy runs on. And brilliant code on a corrupted operating system still crashes. Right now, most organizations are shipping the most disruptive workplace technology in a generation onto an operating system they’ve never patched and never enabled.
I call this the manager-as-algorithm problem. And today I am going to try to convince you it’s the quiet variable that separates the AI rollouts that lift your people from the ones that quietly hollow them out.
Let me give you the numbers, because this isn’t a hunch. It’s some of the better-documented research we have.
Gallup studied people who voluntarily quit their jobs, and they asked a beautifully simple question: could anything have kept you? And the answer should stop every leader who’s betting big on AI. Forty-two percent of those voluntary leavers said yes — their exit was preventable. A manager or a leader could have done something. And didn’t.
Now sit with the reason why. In the three months before they walked out the door, forty-five percent of those leavers said that nobody — not a manager, not a leader, nobody — proactively talked with them about their job satisfaction, their performance, or their future.
Almost half of them left in a silence nobody broke.
And here’s why that matters so much right now. We’re about to drop AI anxiety into a workforce where half of all departures are already preventable and already happening in silence. That’s pouring accelerant on a fire you can’t see yet. Because people don’t quit a chatbot. They quit the feeling that nobody is steering them through the change.
But the same research holds the way out. Gallup found that when a manager has just one meaningful conversation a week with each direct report, employees are four times as likely to be highly engaged. Frontline, hybrid, remote — it doesn’t matter. Four times. From a fifteen-to-thirty minute conversation. That is the single highest-leverage, lowest-cost engagement driver in your entire building. And it runs entirely on enabled managers.
And I want to underline something in that finding, because it’s easy to skate past. We’re not talking about a performance review. We’re not talking about a survey, or a pulse check, or another dashboard. We’re talking about a conversation. A human being, sitting across from another human being, asking how the work is going and waiting for the answer. That’s it. The most advanced retention technology available to you this year costs nothing and fits in fifteen minutes — and most organizations can’t reliably make it happen, because they’ve buried their managers so deep in administrative noise that the one thing only a manager can do is the first thing that falls off the calendar.
So, hold those two numbers together. Forty-two percent of quits are preventable. One weekly conversation quadruples engagement. The gap between those two facts is your entire opportunity — and it’s a gap a tool will never close for you.
So, here’s the distinction I really want you to hold onto. Most AI rollouts are an announcement. A tool shows up, an email explains it, a training video gets assigned, and leadership moves on — and then they’re puzzled when adoption stalls and morale dips.
Announcement is not enablement. An announcement informs the manager. Enablement equips the manager to translate.
And that translation layer is expensive. Gallup estimates that replacing a manager runs about two hundred percent of their salary. These are not interchangeable parts. So, when you strand them, or burn them out, or quietly let them disengage, the whole rollout breaks downstream of them — customer by customer.
So, what does it take to enable the algorithm? Three things the announcement always skips.
First: context, before the tool. Managers can’t translate what they don’t understand. Before that copilot lands on Monday, the manager needs the why — what problem this solves, what it will and won’t do, and the honest answer to the question every employee is silently asking, which is “what does this mean for me?”
Second: permission to coach, not just deploy. We hand managers a tool and a target and call it a day. But a manager’s real job in an AI transition is human — notice who’s anxious, name the fear, hold that weekly conversation. And that only happens when you give them the time and the explicit permission to do it, instead of burying them in the very dashboards AI was supposed to simplify.
Third: air cover to be honest. Enabled managers can say hard true things — “yes, this changes your role, and here’s how we’ll grow you into the new one.” Managers who fear for their own standing default to corporate fog. And people see fog instantly.
Let me ground this in a story. Years ago, when I was inside Starbucks working with their culture, the principle I heard over and over was a sequence: take care of the partners — that’s what they call their people — and the partners will take care of the customers. Take care of the people who take care of the people.
But the mechanism behind that sequence was always the store manager. Corporate could set the values, design the apron, write the whole playbook. Whether a barista felt genuinely seen on a brutal Monday morning rush came down entirely to the person running that store.
Enable that manager, and a national brand promise comes alive at 7:14 in the morning. Strand that manager, and the exact same playbook produces a tired, transactional shrug.
So, here’s the translation. Your AI model is the brand promise. Your manager is the store. And no promise survives a stranded manager.
So three moves, and then I’ll let you go.
One: brief your managers first, and differently. Give them a head start — the strategy, the honest talking points, the real answer to “what about my job?” A manager who hears the news at the same moment as their team has nothing to translate. They become a bystander to their own rollout.
Two: protect the weekly conversation like it’s revenue. If fifteen minutes quadruples the odds of high engagement, defend it from the meeting tide.
Three: measure the manager, not just the tool. Your adoption stats will lie to you about engagement. Ask instead — are the conversations happening? Do people feel equipped, or surveilled?
And here’s the reframe I’ll leave with you. Stop asking “which algorithm should we deploy?” Start asking “have we enabled the algorithm we already have?” — the human one, standing in front of the team, deciding in real time whether your AI feels like a gift or a threat.
You’ve invested fortunes in the algorithm made of code. So, here’s my question: what have you invested in the algorithm made of people?
If you’re not sure your managers are equipped to turn AI into engagement instead of anxiety, that’s exactly the work I do with leadership teams. Please reach out to me at josephmichelli dot com slash contact.
Thanks for listening. I’ll see you next time.
