The AI tools healthcare teams actually use

Healthcare has the widest gap between demand and supply of any market we track: an average opportunity score of 82 against a saturation score of 11. It also has the fewest thin products of any market in the catalog. Those two facts are the same fact, and this is what the 137 live products on the other side of that gap are actually doing.

Why this market is still open

The barrier in healthcare is not intelligence. A general-purpose model can already summarise a consultation, draft a referral letter, or read a scan competently enough to be interesting. What it cannot do is survive the rest of the job: the compliance review, the integration with a hospital’s record system, the procurement cycle, and a buyer who is personally and professionally liable if the output is wrong.

That is why almost no healthcare product in the catalog is a thin layer over somebody else’s model — the market’s average wrapper-risk score is 8.8, third lowest of all 29. It is also why the market has stayed uncrowded while writing tools multiplied. The work that makes a healthcare product real is slow, unglamorous and expensive, and it cannot be skipped.

What the leading products actually do

Sorted by opportunity score, the top of this market clusters into four jobs: taking the documentation burden off clinicians, triaging urgent cases faster than a human queue, automating the insurance layer, and accelerating discovery research. Each of those has a buyer with a budget and a number they can point at.

Note what is absent. There is no general-purpose “medical chatbot” near the top of this market, and that is not an accident of our scoring. Products that answer health questions in a text box compete directly with the free consumer assistants, carry the liability of clinical advice, and have no route to a hospital budget. The products that score well took a narrower job and went deeper into it.

Healthcare barely does free

Only 13 of 119 products with a known pricing model offer a free tier. Across the whole catalog it is closer to two thirds. This is the clearest commercial signal in the market: healthcare software is sold, not adopted bottom-up, and a free tier does almost nothing to reach the person who signs.

For a builder that cuts both ways. You will not get the viral bottom-up adoption curve that a coding tool can get, and your first year will be spent on pilots rather than signups. In exchange, the people who do buy renew on contracts and rarely churn to a competitor over a pricing page.

Who is funding it

Series A dominates at 65 companies. That is a market past the experiment stage and into the part where products have to prove they can be sold repeatedly — which fits what the rest of the data says about long sales cycles and contract revenue.

It is one of the safest markets to be in

Healthcare has lost 8 of 145 tracked products, a 5.5% failure rate against a catalog-wide 8.4%. The markets around it at that end of the table — Security, Learning, Data — share the same trait: a buyer with a line item, not a subscriber with an app store. Our mortality report has the full ranking.

If you are thinking about building here

The honest version is that a wide gap is a description of a barrier, not an invitation. The opportunity score is high because the work is hard, and noticing the opportunity does not make the work shorter. What changes the maths is holding an asset that shortens it — a clinical network that will pilot with you, a dataset with consent already attached, a relationship with a payer, or a co-founder who has been through a compliance review before.

If you hold one of those, this is the market in the catalog where it is worth the most. If you do not, the model you use will not save you, because every product on the list above is using roughly the same one.

The market, product by product

Ranked by opportunity score. Each name links to the product’s own site.

Photo: Tima Miroshnichenko / Pexels. Colour-graded for Falcoscan.

← All articles