June 26, 2026 · Alex, MD
You have the number. Here is the conversation.
The data was the easy part. Asking for it is where most of us freeze. Here is the method, and the line AI cannot cross.
I had AI pull what a night of call is actually worth, and verify it across three tools, so I had a number I could defend. Then I sat with the harder problem. Even with the number in front of me, the ask feels impossible. You do not want to seem greedy. You are talking to someone who negotiates this every day while you do it once every few years. And you are tired. So you take what you are handed. The number was never the hard part. The conversation is.
Start with the data and your goals
You did the data already. The other half is knowing what you actually want before you open your mouth. The number you are aiming at. The number you will walk away at. And the things that are not money that you would happily take instead, a protected day, a cap on consecutive nights, a better schedule. Write all of that down first. A method only works if you know where you are pointing it.
Then borrow a real method, not your instincts
Most of us walk into these conversations and improvise, which is exactly how you end up splitting the difference and calling it a win. So I did not improvise. I used the best framework I know, Chris Voss’s Never Split the Difference. Voss ran hostage negotiations for the FBI, and his whole point is in the title. You do not meet in the middle. You understand the other side’s box, and you trade across different things until both of you get what you actually need.
Here is the part that is new, and the reason this belongs in a piece about AI. Most of Voss is preparation, and preparation is exactly what AI does well. So I used it to implement the method, not just read about it. I had it map the other side’s constraints, draft the labels and the questions, write the line that disarms the objection before it lands, and even play the other side so I had heard the pushback once before it was real. Here is what that looked like, step by step.
Map their box first. The hospital legally must keep call covered, and it cannot pay above fair market value. That constraint is your ground, not your obstacle. Knowing it is half the negotiation, and AI is good at laying it out.
Open with a label, not a demand. “It seems like keeping this call covered has been a real problem.” You name their world before you ask for anything.
Ask a calibrated question instead of stating a number. “How am I supposed to keep taking sixty nights a year at the current rate?” It hands them your problem to solve.
Disarm the objection before they say it. The accusation audit: “You are going to think I am only here about the money.” Said first, it loses its power.
Put down a precise number, not a round one. The data said anesthesia call runs about $1,656 a day, higher at a trauma center. I would anchor near the top of what I could source and say $1,850 a night, not $2,000. Precise reads as researched. Round reads as plucked from the air.
Then trade something that is not dollars. A protected day after a string of nights. A cap on consecutive calls. That is how you avoid splitting the difference on the number itself.
And the move that beats the paywall. “I want this to sit cleanly within fair market value, so which survey and which percentile are we working from?” The real numbers are locked behind a survey neither of you can see for free, so this is not a gotcha. It is the only way either of you gets to an honest number.
I had AI play the other side first, the administrator who says there is no budget, so I had already heard the pushback once before it was real.
The line AI cannot cross
Here is what none of it does. It cannot read the person across the table. It cannot tell you when a fair deal is just good enough to sign, or whether one more protected weekend is worth more to you than the money. It can hand you the labels, the questions, and the number. It cannot sit in the chair. The prep gets you to the table ready. The room is yours.
The takeaway
You stop walking in to find out what your time is worth and start walking in already knowing. The data does the watching. The judgment, when to push, when to trade, when to stop, stays yours. That is what a little breathing room looks like. Not a number a machine handed you, but the calm of knowing what you can defend and asking for it like you mean it.