AI Can’t Drive a Candidate’s Partner Around a Neighborhood

Let me be fair about what AI has done for physician recruitment, because it’s a lot, and pretending otherwise makes the rest of this argument unserious.

It has made the process from sourcing to hire legible. You can track conversion. You can see where the bottleneck actually is instead of guessing. It handles the note-taking, condenses information, moves it between team members without anyone retyping it at 9pm. It surfaces lists of candidates who are actually in motion. All of that is real, and none of it should go back.

So the objection isn’t that AI is in the room. It’s what we do with the hours it hands back to us.

The line

The line shows up at a specific moment, and it’s the same moment every time.

It’s when a person wants the thoughtful insight, the experience, and the truth of another human being. When they want to feel what the people at this organization feel. When they want to understand, early, that someone here sees them as a person and not a fill.

We are starved for that. Not metaphorically. People are starved for the experience of being heard and seen, and when you can offer it without an expectation of return, with an honest wish for what’s best for them, it registers. Immediately. It’s the most noticeable thing you can do.

Sometimes what’s best for them isn’t your organization. That’s fine. If you’ve actually listened, you’ll know what kind of place fits, and you’ll probably know someone there. That’s not a lost recruit. That’s a person who will describe you accurately for the rest of their career.

What breaks

Relocation isn’t information. It’s a personal experience, and it belongs to a specific family with a specific spirit that AI cannot reasonably know well enough to advise.

So the one-size-fits-all community section rings hollow, and worse, it reads as an afterthought. Which is a message. It says: we care what you do inside this building. What happens to you outside of it is a PDF.

Here’s a realtor. Their profile says relocation on it.

The stories are consistent enough to be a pattern:

A mid-career physician says plainly that they want to be downtown, in the heart of the city. They get driven to the suburbs anyway, because the person touring them decided an affluent buyer belongs in an affluent suburb and downtown is “dangerous.” Nobody asked. The candidate stated a preference and was overruled by an assumption.

A physician loses their dream home. Then another one. Their offers keep getting passed over because their agent doesn’t know how to present a physician loan as strong in a competitive situation. The buyer was qualified the whole time. The representation wasn’t.

A rural state sends its relocation guide, and the guide is all city amenities — the nearest metro, the airport, the restaurants an hour away. The candidate wanted the quiet. Wanted the slower version. Never saw it, because the guide was written by someone who assumed everyone wants the city and was apologizing for the absence of one.

Every one of those is a canned answer meeting a real person and losing.

Nowhere is this worse than school selection

Ask a relocating family what matters most and school will be in the top two. Then watch how that question gets answered.

It gets answered with a website. A number out of ten. A ranking. Or the phrase that does the most damage of any phrase in relocation: the best school district.

Best for whom.

Niche and its cousins are aggregating test scores, college placement, and reviews from strangers. That’s a real signal about something. It is not a signal about your family. It cannot tell you whether your kid — who has an IEP, or is autistic, or is bored senseless in a lecture, or plays cello four hours a day, or is anxious in a class of thirty-two — will be okay in that building. It doesn’t know if you want your children in a school that looks like the world, or a school where a teacher will notice on day three that something’s off. It can’t tell you which principal actually returns calls. It doesn’t know that the “10” district has a culture your family would find suffocating, and that the “7” a mile away has the arts program that would change your daughter’s life.

A ranking measures what’s measurable. A family is deciding on what isn’t.

This is what a real school consultant does that a website structurally cannot: they ask what you value first. Then they translate. They know the difference between the school’s reputation and the school’s actual environment, because they know families in it. They know which programs exist on paper and which ones exist. They know which schools handle a mid-year arrival gracefully and which ones will leave your seventh grader eating lunch alone until March. They know about the waitlist that isn’t really a waitlist and the charter that’s worth the drive and the magnet whose application closed in January.

And they know what a doctor’s schedule can do to a family. That’s a variable no ranking contains. Which schools will work when one parent is unreachable for twenty-eight hours at a stretch. What after-care actually looks like. Whether a eleven-year-old can get home when both parents are working and one of them is on call.

Handing a family a link to a rankings site is not answering the question. It’s declining to.

What isn’t in the dataset

AI cannot tell you which coffee shop the physicians actually go to. Or the brunch spot. Or which brewery, on which night, because it’s a standing thing and everyone just knows.

It cannot tell you that the spouses meet at that park on Wednesdays and bring the kids.

It cannot tell you the block-by-block variables inside a neighborhood — that this side of the street is different from that one, and why.

It cannot tell you what the drive is actually like from this address to that hospital at 6:40 on a Tuesday morning in February. Not the average. The real one.

And it cannot tell you what it’s like to work inside these particular hospital systems, or to be married to someone who does, or how each system actually treats the people who work for it.

That knowledge exists. It’s just that the only way to have it is to have been there, and the only way to get it is to ask someone who was.

The actual threat

The threat was never AI.

The threat is trusting AI with the things that require a human, and then forgetting to put the human back in.

Efficiency is not the goal. Efficiency is the thing that buys the goal. AI exists to give you time back — face to face, voice to voice, with the candidate and the family who are about to reorganize their entire lives around a decision you helped them make.

Use it to give you time back to be human.

Then go be human.