Your best person just quit. Did their AI workflow quit with them?
When your best employee leaves, you expect to lose their talent. What most leaders don't expect to lose is the AI workflow they built quietly on their own — the one that was cutting six hours of work down to forty-five minutes every week. Nobody wrote it down. Nobody else knows how it works. And in fourteen days, it walks out the door with them.
Cloud Systems Engineer · MCT · PMP · Azure Solutions Architect Expert. 40+ years in IT and operations. Builder of the Blair AI Rollout Framework.
Published July 2026
Mary is VP of Operations at a 400-person logistics company. Her best ops lead just gave notice — and took an AI workflow cutting six hours of reporting down to forty-five minutes with her. Nobody wrote it down. Today we talk about what to do in the next two weeks, and how to build the process so it never happens the same way twice.
Listen on SpotifyThe resignation letter you didn't see coming.
It's a Tuesday morning. Your best ops lead — the one who runs weekly reporting for the entire division — just gave two weeks' notice. That's hard enough. But here's what makes this one different.
Somewhere along the way, without being asked, without any announcement, she built a set of AI prompts that cut her reporting time from six hours a week down to forty-five minutes. She figured it out on her own. She kept using it. And nobody wrote it down.
In fourteen days, that workflow walks out the door with her. And nobody will even notice it's gone — until the reporting that used to take forty-five minutes starts taking six hours again, and nobody can explain why.
"Organizational AI memory is what separates a team that grows its capability over time from a team that starts over every time someone leaves."
1. You didn't lose an employee. You lost institutional knowledge you didn't know you had.
When AI adoption happens informally — one person experimenting on their own, quietly getting faster — the knowledge doesn't live in the organization. It lives in that person's head. And anything that only lives in one person's head is one resignation letter away from disappearing.
Here's what makes this different from a normal knowledge transfer problem. When someone leaves and takes a traditional process with them, you usually know what you're losing. You can see the gap. You can plan around it.
But when someone leaves and takes an AI workflow with them — one that was never written down, never shared, never officially part of how the team operates — you don't always know what you're losing until it's already gone. The reporting that used to take six hours is suddenly taking six hours again. And nobody can explain why.
Informal AI adoption is invisible until it isn't. When it becomes load-bearing — when the organization depends on it without knowing it — and then the person who built it leaves, you lose capability you didn't even know you had.
2. This isn't a people problem. It's a systems problem.
Most leaders don't find out how dependent they've become on someone's shadow AI work until that person is already gone. Not because they weren't paying attention — but because informal AI adoption is invisible by design. Nobody announced it. Nobody documented it. It just quietly became load-bearing.
Think about what actually happened. The ops lead found a problem — six hours of reporting every week that felt like it shouldn't take that long. She experimented. She built something that worked. And then she just kept using it. That's not negligence. That's initiative.
The problem isn't that she built it. The problem is that the organization never created a way to capture what she built. No documentation process, no workflow library, no expectation that when someone finds a better way to do something with AI, that knowledge gets shared and owned by the team — not just the individual.
And that's not her fault either. She wasn't asked to document it. She wasn't given a reason to. This is a leadership and systems problem — not a people problem.
3. You still have two weeks. Here's what to do with them.
The fix isn't to stop people from building AI workflows on their own. That instinct — finding tools that make work faster, experimenting, figuring things out — is exactly what you want in a team. The fix is to build a culture where those workflows get captured, documented, and owned by the organization rather than the individual who built them.
But first, use the window you have.
Sit down with your ops lead before she leaves.
Not to guilt her. Not to make her feel responsible for a gap she didn't create. Ask her one thing: can you walk me through exactly how you do this? Record the conversation. Take notes. Ask her to write down the prompts she uses, the steps she follows, the decisions she makes along the way. You probably won't capture everything. But you'll capture enough to keep the lights on — and enough to build a real documentation process so this never happens the same way twice.
Build the process that prevents it next time.
Create a simple workflow library — even a shared document — where employees are expected to document any AI process they build that affects team output. Set the expectation clearly: when you find a better way to do something with AI, write it down and share it. That expectation, built into how the team operates, is what transforms individual AI knowledge into organizational AI memory. The goal isn't to slow down your best people. It's to make sure that when they leave, they take their talent with them — not your process.
This week's AI Hot Tip.
Identify one person on your team who you suspect has built something with AI that nobody else knows about. Could be a reporting shortcut, a prompt they use to summarize meeting notes, a template that speeds up a weekly handoff. Ask them one question: "Can you walk me through how you do that?" Then write it down — even a rough paragraph, even a voice memo you transcribe later. That five-minute conversation is the beginning of organizational AI memory.
Three things to take with you.
- Informal AI adoption is invisible until it isn't. When someone builds a workflow on their own and never documents it, the organization doesn't own it — that person does. When they leave, it leaves with them.
- This isn't a people problem. It's a systems problem. Your team's instinct to experiment and find faster ways to work is exactly what you want. What you need to build around that instinct is a way to capture what they find.
- If you're in this situation right now, you still have two weeks. Use them. Sit down with your ops lead, record the conversation, write down the prompts. You won't get everything — but you'll get enough to build from. Then build the process that means you never lose it the same way twice.
Find out if your AI capability lives in your people or in your process.
The AI Readiness Score gives you a documented baseline of exactly where your organization stands today — including whether the AI knowledge your team is building is being captured or walking out the door. About five minutes, no email required to begin.
Start Here — Free AI Readiness Score →Related resources.
Shadow AI Guide →
What to do when AI is already in your organization before you rolled it out.
I Don't Know Enough About AI to Lead →
You don't need to be the expert. You need to be the structure.
Your First 30 Days with AI →
The complete 30-day plan before you touch a single tool.
AI Workforce Communication →
How to talk to your employees about AI before the rumor mill does.
Common questions.
Does your AI capability live in your people or in your process?
The free AI Readiness Score gives you a documented baseline in about five minutes — including whether the AI knowledge your team is building is actually being captured or quietly walking out the door.