When the Algorithm Reads Your Chart Before Your Doctor Does

You send a message to your clinic portal at 11:40 on a Tuesday night. Something about the spotting. You're not sure if it matters. A reply comes back in ninety seconds.

That speed used to mean a nurse was awake and worried enough to answer. Now it might mean something else entirely.

AI is making its way into fertility care in some of the same ways it has entered other corners of medicine — not necessarily in the embryology lab first, but potentially in the scheduling system, the benefits portal, the chatbot that asks what you're calling about before it decides who gets your call. Patients aren't always told when this is happening. You might just notice that some things got faster and some things got stranger.

So it's worth knowing what's actually happening, and what it means for you.

The argument for it is better than you'd expect

The pitch from inside the industry isn't that AI will diagnose you. It's that AI can absorb the enormous administrative sludge that currently eats the time your care team could be spending with you. Progyny's product leadership has made the case that AI in healthcare can make care more human rather than less — if it's designed around what patients actually need rather than around cutting costs, with the promise concentrated in navigation, personalized guidance, and reduced administrative burden.¹

That framing deserves a fair hearing, because anyone who's been through a fertility cycle knows how much of it is logistics. Prior authorizations. Pharmacy transfers. Figuring out whether your plan covers the genetic testing or only the biopsy. Calling three numbers to find out which one handles your case. None of that is medicine. All of it is exhausting, and most of it lands on you at exactly the moment you have the least capacity to deal with it.

If a system can tell you in ten seconds what your remaining benefit balance is, instead of you sitting on hold for forty minutes to learn the same thing, that's not a threat to the human part of care. That's a defense of it.

Where AI could genuinely be useful to you

Understanding your coverage. Fertility benefits can be difficult to navigate — lifetime maximums, cycle limits, carve-outs, medication covered separately from procedures. A well-built navigation tool could, in principle, hold all of that at once in a way a stressed human on the phone might struggle to.

Appointment preparation. If a tool can pull your recent labs, your cycle history, and your last visit notes into a summary you can actually read before a consult, that could help you walk into that room a different patient — asking better questions, rather than discovering three days later that you forgot the one thing you meant to ask.

Tracking and pattern-spotting. Medication timing, symptom logs, injection schedules. This is the kind of repetitive, high-stakes detail work where a system designed for it could outperform a person who is grieving and sleep-deprived.

Translation. Turning "endometrial thickness 6.8mm, trilaminar" into a sentence in plain English. Not a diagnosis, not a prediction. Just a translation. That could be a real service, especially if you're the person Googling your own ultrasound report at 2am.

Where it can't go

Here's the line I'd hold.

AI is good at retrieval, sorting, and summarizing. It is not good at judgment under uncertainty with incomplete information and a person's life on the other side of the decision — which is a fairly precise description of most of the hard moments in fertility care.

Whether to convert a cycle. Whether to transfer the single euploid or wait for another retrieval. Whether this is the point where you stop. These decisions aren't limited by information. They're limited by values, risk tolerance, finances, exhaustion, and how much more of this a particular person can carry. A model doesn't know what you can carry. It doesn't know that you've already decided this is the last one and you haven't told anyone yet.

The same is true of the emotional weight. There is no version of a chatbot that can sit with you after a failed transfer in the way a nurse who has known you for two years can. Some of what makes care bearable is being known — that someone remembers your name without looking it up, remembers this is your third cycle, remembers what happened last time. That's not a feature. It's a relationship, and it accumulates slowly.

Be alert to the difference between AI that gives your team more time with you and AI that replaces time with you. Both look like efficiency on a spreadsheet. They feel nothing alike from the exam table.

Questions worth asking

You don't need to become a technology skeptic. You do get to be a curious patient. A few things worth raising, without apology:

"Is a person reviewing this before it reaches me?" Especially for anything involving results, medication instructions, or a change to your protocol. You're entitled to know whether a human has laid eyes on it.

"Who sees my data, and where does it go?" Fertility data is unusually intimate — cycle dates, pregnancy outcomes, genetic results, sometimes information about a partner or donor. Ask whether it's used to train anything, whether it's shared with your employer's benefit administrator, and how long it's retained.

"If I'd rather talk to a person, how do I do that?" There should be a clear answer. If there isn't one, or if getting to a human requires four steps and a phone tree, that tells you what the system is actually optimized for.

"How does this tool handle cases like mine?" Models are built on aggregate data. If your situation is atypical — unexplained infertility, recurrent loss, a rare diagnosis, a same-sex couple whose path doesn't fit the default template — it's fair to ask how much of what you're being shown was built with someone like you in mind.

The part that stays the same

Technology in fertility care has always been a mixed blessing. Every advance has arrived with a promise and a cost, and patients have always been the ones absorbing the gap between the two. AI won't be different.

What helps is remembering that the tools are yours to use, not the other way around. Use the portal that tells you your benefit balance. Use the summary that helps you prepare. Then close the laptop and ask the human being in the room the question you actually came to ask.

The most valuable thing in fertility care has never been information. It's been having someone in your corner who knows the whole story. That hasn't changed, and no one should let you believe it has.

¹: Progyny, "AI Can Make Healthcare More Human—If We Design It That Way"