He picked the one corner of medicine everybody avoids - the prior authorization form - and decided that was exactly where the interesting work was hiding.
Prior authorization is the part of American healthcare nobody puts on a brochure. A doctor orders a test. An insurer wants paperwork first. Someone spends an afternoon transcribing policy rules onto a form, the patient waits, and the whole thing crawls forward at the speed of a fax tone. It is tedious, expensive, and quietly one of the largest sources of friction in the system. Amber Nigam looked at that and saw a place to build a company.
basys.ai, the company he co-founded with Jie Sun, takes the rulebook insurers use - dense, conditional, constantly changing - and turns it into something software can read. The engine was trained on more than 10 million patient records and claims. The pitch is blunt: automate up to 90% of prior authorization requests, hit savings targets months sooner, and give clinicians back the hours they were never supposed to spend on forms.
The work sits at an awkward intersection. Payers want cost control. Providers want speed. Patients want a yes. Those incentives rarely point the same direction, which is the entire reason the problem has survived this long. Nigam's bet is that explainable, auditable AI can serve all three at once - not by picking a side, but by making the process fast and transparent enough that the fight loses its oxygen.
It is also a deeply unglamorous thesis, and that is the point. There is no consumer app, no glossy wearable, no viral moment. There is policy, encoded. There are audit trails. There is the unsexy machinery of getting a treatment approved before a patient gives up waiting. Nigam talks about it the way some people talk about cathedrals - the value is in the parts you do not see.
He did not arrive at this from the outside. While building basys.ai he was finishing a master's in Health Data Science at Harvard, and moonlighting as an associate director of the Harvard GSAS Business Club, where his job was helping other founders sharpen their pitches. He has been on both sides of the table - the one asking for money and the one teaching people how to ask.
basys.ai was born inside Harvard in 2021-2022 and grew up in Cambridge, the kind of zip code where a healthcare startup can borrow credibility from the institutions next door. By 2023 the company had closed an oversubscribed round totaling $2.4 million, with a roster of backers - Eli Lilly, Mayo Clinic, Nina Capital - that reads less like a seed deck and more like a vote of confidence from the industry it wants to fix.
The product is generative AI, but the magic is in the plumbing - the encoding, the matching, the paper trail nobody usually wants to look at.
Dense, ever-changing payer policy gets translated into structured logic a machine can actually reason over.
An authorization request is checked against the encoded policy in real time, surfacing what is needed and why.
Decisions arrive with audit trails and evidence documentation, so approvals are explainable - not a black box.
Figures per basys.ai public materials and 2023 funding announcement.
Work appearing at top AI venues including NeurIPS and ACL, and with publishers including Springer and Lancet. Roughly a dozen peer-reviewed papers cited.
First inventor on three patents - the rare CEO who can claim the IP as well as pitch it.
Boston Congress of Public Health (2022) and Boston Business Journal (2025), plus a Top 50 in Digital Health nod from Rock Health.
Selected for prestigious founder fellowships supporting work in social innovation and health.
Has taken the mic at TEDx on aligning healthcare incentives, and presented at the Mayo Clinic, Harvard, and MIT.
Writes on AI bias, trustworthy AI in patient care, prior authorization reform, and leadership authenticity.