As experienced leaders leave, the decisions their rules don't cover will go to managers who have never made them. Those managers need to practice first.
On January 15, 2009, about two minutes after US Airways Flight 1549 left LaGuardia, the captain said one word: "birds." About a second later, the birds hit. Both engines lost almost all their thrust. The airplane was 2,818 feet above the ground.
About 12 seconds after the birds hit, the captain took control of the airplane. Five seconds later, he told the first officer to get the checklist for losing thrust on both engines.
That checklist, the National Transportation Safety Board (NTSB) later wrote, was "designed assuming that a dual-engine failure occurred at a high altitude (above 20,000 feet)." Flight 1549 was more than 17,000 feet below that.
The checklist was three pages long. The crew completed most of the first part. They never reached the section on landing in water.
The biggest decision was not on those pages. The captain chose the Hudson River instead of trying for an airport. That choice, the NTSB concluded, "provided the highest probability that the accident would be survivable." All 150 passengers and five crew members left the airplane.
The checklist was not bad. It was the most applicable one the crew had, but it "did not fully apply" to this emergency, and where it did not apply, the crew decided.
Companies can now write their own version of that checklist, with AI. The checklist will stay. The person who knew when it did not apply may not.
The People Who Knew When the Rules Didn't Apply Are Leaving
AI can now write down how your senior leaders decide, but only for the cases they expected. As experienced leaders leave, the managers who stay will make the decisions those rules don't cover. They have to practice those decisions before they become responsible for them.
AI that writes down how leaders decide is already here. In a McKinsey article published November 2, Kurt Strovink and Sangeeth Ram write that "AI can now codify that tacit wisdom and make it available to everyone, for the first time." They also name a risk: "A system tuned to one leader's standard can hand a successor dogma. The record must hold the conditions and reasoning behind a standard, not only its conclusions."
The departures are starting. Microsoft is offering people ways to leave: it introduced a Voluntary Exit Offer on October 26, after a Voluntary Retirement Program earlier this year.
In October we wrote that AI leaves your managers fewer decisions, and more difficult ones. The person who used to make the most difficult decisions may not be there to ask.
Here is one sales team. A regional sales manager takes over the pricing-exception decisions of a vice president who is leaving. The vice president's rules are now part of the team's AI deal assistant. The first difficult renewal brings three decisions.
Decision 1: The Rule Says Keep the Price. Is That Still True?
The common move: Follow the assistant. It applies the vice president's rule, and the rule says this customer's one-year renewal gets the standard price. The manager keeps it.
The cost: The vice president wrote that rule when no competitor was involved. This year, a competitor has made the customer an offer. The customer leaves, and the records show the manager followed the rule.
The move that works: Before you apply a rule, list what it assumed: which competitors are involved, the contract term, the customer's history. Then check which of those are still true. Here, one is not.
Decision 2: The Rule Doesn't Fit. Who Decides?
The common move: Once the manager sees that the rule doesn't fit, the manager refers the renewal to the new vice president, who has never met the customer, and waits.
The cost: The answer arrives after the customer's deadline, or it comes from someone who knows less about the customer than the manager does.
The move that works: Decide within your authority. Say what you decided, what fact would make you decide differently, and when you will review the result. If the decision is outside your authority, refer it with your recommendation, not only the question.
Decision 3: You Lowered the Price. Now Tell the Team Why.
The common move: Lower the price and say nothing.
The cost: The team sees an exception to the rule with no explanation, so it learns nothing about when the rule applies. When you leave, nobody else knows the reason.
The move that works: Tell the team which condition changed and why that changed the decision: "The rule assumed no competitor. There is one now, so I lowered the price this time." Now the team knows the reason, not only you.
Why Better Rules Aren't Enough
Those three moves are easy to read. They are difficult to make for the first time, with a customer waiting.
McKinsey's own advice is a better record: the conditions and reasoning, not only the conclusions. So add the competitor condition, update the assistant, write a handover document. Do all of that. But each new rule still covers only the cases someone expected. Someone still has to notice, at the time, that a condition is no longer true. The people who were good at noticing are leaving, so a manager's first attempt may be with a real customer. Knowing the rule is not the same as making the decision.
A simulation lets that first attempt happen before a real customer is involved. It shows what a manager does after a competitor is mentioned briefly, in a conversation about something else.
A Lost Renewal Looks Like the Market's Fault
If nobody practices, the losses are hard to see. Count the decisions a departing leader made in a year: exceptions approved, customers kept, risks accepted or refused. After the leader leaves, each depends on a manager's first attempt. When one goes wrong and the records show the manager followed the rule, the loss looks like the market's fault. Nobody changes how the decision is made.
Practice the Decision Before a Customer Depends on It
Blueline starts with one of those decisions. Decision-Consequence Mapping™ (DCM™) maps a decision a departing leader has been making, starting from the business number the decision affects, such as renewal margin. The people who know that decision best help us map it. Together we list the conditions the rule depends on and, for each change, the common move, the move that works, and the consequence of each.
In a simulation built from that map, a manager is responsible for a renewal that the rule says should get the standard price. The manager's account executive briefly mentions that a competitor has made an offer to the customer's head of purchasing. Nobody says that the pricing rule no longer applies.
The manager can keep the standard price, refer the decision to the new vice president and wait, or decide and explain. Keeping the standard price loses the renewal. Referring it costs time, and the head of purchasing becomes less willing to work with the manager. Deciding the price with the competitor's offer in view keeps the customer.
What lets every manager on your team practice that decision is BluEQ™, Blueline's proprietary behavioral intelligence infrastructure. It carries the DCM™ design into every conversation, so each character remembers what the learner did and responds the way the real person would.
Before any real renewal depends on it, you see each manager's decision path: who kept the standard price, who referred the decision, who decided and explained why, and at which point. You name the metric and provide access to the data that measures it. Its current value becomes the baseline for ROLI (Return on Learning Investment).
Start Here: What Does Your Rule Assume?
Choose one decision a departing leader makes today, and the rule the team will use after that leader leaves. Ask three questions:
- What does the rule assume about the situation? Write the conditions down.
- When one of those conditions is not true, who decides, and with what authority?
- When someone makes an exception to the rule, who hears why?
If nobody can answer the first question, your rules record decisions without the conditions behind them.
Then describe that decision to us in a private demo, and we'll talk about what a simulation of it would look like. If someone else is responsible for the results of that decision, send them this post and the ROLI Blueprint, which explains the business case.
After the Hudson, the NTSB Asked for a Better Checklist and More Practice
Among its recommendations after Flight 1549, the NTSB asked for a checklist "for a dual-engine failure occurring at a low altitude." It also asked that pilots' ground and simulator training include that scenario, which, the Board found, "would challenge them to use critical thinking."
The Board did not choose between them. Your managers need both, and the practice has to happen before they become responsible for those decisions.
Sources: NTSB, Aircraft Accident Report NTSB/AAR-10/03, US Airways Flight 1549 (adopted May 4, 2010): summary and recommendations A-10-66 and A-10-69; full report, section 1.1; findings 12, 15 and 21. NTSB, Safety Recommendation letter A-10-88 through -95 (May 21, 2010). Kurt Strovink and Sangeeth Ram, "Turning executive judgment into organizational intelligence: Leveraging AI in CEO leadership," McKinsey & Company, November 2, 2026. Amy Coleman, "Introducing a Voluntary Exit Offer," Microsoft, October 26, 2026.
