“Bring Me Solutions”
Puts the Fix Before the Problem.

Missed appointments at your clinic have been climbing since spring, so your clinic manager spends a weekend using AI to build a working reminder app. On their phone, they show you the text a patient would get the day before, with a button to confirm. The app looks finished, and you and your own boss approve it that week. Two months later, missed appointments have barely moved.

Nearly all of the rise is in the 7 a.m. slots your clinic added in spring, and the first bus gets there after seven. One of your nurses noticed who was missing at seven and never raised it, because the only question anyone asked was when the app could start. When you ask your team for solutions first, they can end up solving only their first guess at the problem.

You have probably said “Bring me solutions, not problems,” or had it said to you. It stands for a sound habit: work out the details yourself, and bring one finished answer up for approval. The habit was built to keep half-finished ideas away from the person who decides, and to teach people to think for themselves. The advice means well, but in the rush to a quick solution, people can skip a skill by accident: working out what the problem is, and what outcome would show it is solved.

You need that skill when a problem has many parts that affect each other, or causes that are still unclear. Your missed appointments were like that. The app was a reasonable fix, and reminders do help more patients keep their appointments. But nobody had checked whether patients were actually forgetting.

The advice never tells your clinic manager what you think the problem is, or which fix you will approve, so they guess. They might sense what you want from your tone, or piece it together from the plans you approved before.

So they are doing two jobs at once: working the problem, and predicting what you will sign. That second job adds cognitive load, and it pulls attention away from the missed appointments themselves. Over time, they may learn to describe a problem so it fits a fix they already have, because that is what gets approved.

With AI, a fix like the reminder app can now arrive already built and running, before anyone has asked what is going wrong. I think a working app is harder to set aside than a plan on paper, so you are less likely to ask what else the problem could be.

I am autistic. Noticing patterns and coming at a problem from a different angle is how my mind works. Earlier in my career, I thrived with leaders who saw the value in that. Other leaders found the same thinking a burden, and saw it as a sign I was not executing or as a challenge to views they had long held. I read the difference as a design question: what a leader asks for determines whether that kind of thinking reaches the decision.

Your organization already pays your clinic manager and your nurses to think problems through. Ask for the fix alone, and you get only the last step of that thinking. Sometimes one of them has already found a better problem to solve.

Now that AI makes a quick solution easy to get, working out what the problem is, and what outcome you need, has become the critical skill. When someone brings you a fix, even one that already runs, set it aside and ask these three questions first, in this order. Your clinic manager might answer like this:

Then ask what your clinic manager would do: “Move patients who come by bus to later slots, offer 7 a.m. to patients who drive, and send the reminder texts to everyone.” They still own the problem and the recommendation, and they no longer have to guess what you had in mind. The person closest to the work lays out the facts, then recommends, and you make the call. That order comes from my Decision Charter, the framework I use to decide who does which part of a decision.

I would hand AI the slow part of working out a problem: listing what could be driving it, and drafting other ways to read it. Your people check each of those against what they see, and decide which ones matter. You can run the Operating Prompt below on the next fix someone brings you.

Once the problem comes before the fix, the person who sees it differently gets heard. Before the app starts, you learn what your nurses noticed. Your clinic manager still brings you a plan, and it begins with the patients who could not get there by seven.

Below is a problem someone brought me, with the fix they proposed or built. First, restate the problem in one sentence without the fix, then say what it is costing and what outcome would show it is solved. List what could be driving it, and mark what changed recently. Draft two other ways to read the problem, each from a different point of view, such as the customer or the front line. For each, say what would have to be true and what we would see if it were right. Under “Questions to ask them”, list what I should ask the person who brought it. Only then, say which readings the fix addresses and which it leaves out. Do not choose a reading or make the call. Label anything you infer as a guess, and list any fact you need from me under “Questions for me”. Problem and fix: [paste here]