When Your Clinic Says 'AI-Assisted,' What Should You Actually Ask?

You're sitting in the consult room and the doctor mentions, almost in passing, that the clinic uses AI to help select embryos. Or to time your trigger shot. Or to flag which protocol you're most likely to respond to. You nod. It sounds reassuring — smart, even. Then you get to the parking lot and realize you have no idea what that actually means for your cycle.

This is the moment to slow down.

AI is moving into fertility care fast, and not always loudly. It's in embryo grading software, in cycle monitoring tools, in the benefits platforms your employer uses to route you to clinics, in the chatbots that answer your portal messages at 11pm. Some of it is genuinely useful. Some of it is administrative efficiency wearing a clinical costume. As a patient, you're allowed — encouraged, actually — to know which is which.

The honest tension nobody's hiding

Here's something worth knowing: even the people building these tools are openly wrestling with where AI helps and where it gets in the way. Leaders at Progyny, one of the largest fertility benefits managers in the US, have publicly acknowledged that AI's data and pattern-recognition capabilities have to be balanced against the parts of fertility care that can't be automated — emotional attunement, clinical judgment in ambiguous cases, the trust built between a patient and a care team.¹²

That's a more honest framing than the marketing usually allows. And it's a useful one for you, because it gives you permission to ask: in my care, where is AI doing the thing it's good at, and where is it being asked to do something it isn't?

What AI is actually good at in fertility care

A few things, genuinely:

• Pattern recognition across huge datasets. AI can look at thousands of cycles and spot correlations a single clinician working from memory can't. That can inform protocol selection, predict response to stimulation, and flag patients who might benefit from a different approach.²
• Image analysis. Embryo morphology grading has historically been subjective — two embryologists can rank the same embryo differently. AI-assisted grading can add consistency.
• Administrative load. Drafting notes, summarizing records, handling scheduling logistics. When this works, it can free clinicians to spend more of their visit actually looking at you.¹

None of this is small. If your clinic is using AI well, your care can genuinely be more personalized and more consistent than it would have been ten years ago.

What AI cannot do — and shouldn't be asked to

The harder list:

• Sit with you when a transfer fails. No tool can read the silence in a room.
• Weigh values. Do you want to transfer one embryo or two? Bank more eggs or move to transfer? These are not optimization problems. They're life questions, and they require a clinician who knows you.²
• Catch the thing that doesn't fit the pattern. AI is excellent at the average case. Fertility patients are frequently not the average case.
• Build trust. Trust comes from being heard. A model cannot hear you.

The risk isn't that AI replaces your doctor. The risk is subtler: that AI quietly shifts what your doctor spends time on. If a tool drafts the note, summarizes your chart, and proposes the protocol, your physician may show up to the visit having spent less time actually thinking about you specifically. Efficiency gained somewhere is time spent — or not spent — somewhere else.

Questions worth asking your clinic

You don't need to be combative. You need to be curious. A clinic that's using AI thoughtfully will welcome these. A clinic that's using AI as a throughput tool will get a little squirmy, and that itself is information.

1. "Where in my care are you using AI or algorithmic tools?" Embryo selection? Protocol recommendation? Note-taking? Patient messaging? Get specific.
2. "Who reviews the AI's output before it affects a decision about me?" The answer should be a named human with relevant credentials, not "the system handles it."
3. "What does the AI do when my case doesn't fit its training data?" If you're older, have a rare diagnosis, are doing reciprocal IVF, or are using a donor or surrogate, the model may have seen fewer cases like yours. How is that accounted for?
4. "If I disagree with an AI-influenced recommendation, what's the process?" You want to hear that the clinician has discretion and is willing to use it.
5. "How does AI change how much time my doctor actually spends on my case?" This is the question most clinics aren't asked. Ask it.

Questions worth asking your benefits provider

If you have fertility coverage through an employer-sponsored program, AI is likely shaping your experience before you ever pick a clinic — in how you're matched to providers, in the educational content you're shown, in how your care advocate is supported.¹

• "Is my care advocate a person, an AI, or a person assisted by AI?" All three exist. You deserve to know.
• "How are clinics in your network evaluated? Is AI involved in those rankings?"
• "If I want to talk to a human, how fast can that happen?" Especially after hours. Especially after bad news.

The reframe

Here's the thing it took the industry a while to admit: more efficient is not the same as more human. Sometimes they pull in the same direction — a clinician freed from paperwork has more attention for you. Sometimes they pull apart — a system optimized for cycle volume can quietly stop noticing that you, specifically, are exhausted and need a break.¹²

You are not a data point in a training set. You're a person trying to build a family, and the technology around your care should be earning its place by making your experience better, not just smoother for the clinic.

If AI in your fertility care is doing its job, you probably won't notice it much. You'll notice that your doctor seems prepared. That your protocol seems thought-through. That when something unusual comes up, someone calls you and explains it in plain language. That's the bar. Anything less, and the technology is working for someone other than you.

Next appointment, bring one of those questions. See what happens.