The Doctor's Note Drives Billing and Insurance
That short write-up after a visit becomes the code for billing and insurance, which explains why AI went after paperwork first.
Lightspeed Venture Partners · How AI Seeks to Change Every Step of the Patient Journey | LightworkWhen a visit ends, the doctor writes a clinical note. It summarizes the patient's background, what was found, and what should happen next, in a form the rest of health care can act on.
Lightspeed partner Gllum Iman puts it this way: if the patient journey is the atom of health care, the clinical note is its nucleus. It turns a messy human story into structured, programmatic wording that other parts of the system can use.
That note goes into the medical record, usually an electronic system that stores a patient's history. Then comes medical coding, where companies like Fathom translate clinical language into financial language. A diagnosis and a procedure become standardized codes for billing.
Those codes drive what happens next: claims, insurance review, and payment. That back office work is known as revenue cycle management, and another company, Signal Labs, focuses on how providers and insurers interact so they do not end up in what Iman calls a big AI war with each other.
Iman puts U.S. health spending at roughly $5 trillion, with more than $1 trillion of it administrative.
Iman says artificial intelligence had its most immediate impact on this administrative side, including writing notes and improving claim adjudication. He describes those tasks as already so error prone that a small improvement helps, and as the area AI could hit first, with more complex and higher stakes uses opening up as computing power grew.
Although this work does not directly diagnose or treat patients, Iman argues that lives are still affected, which is why he calls the administrative side a very important space.
Questions people ask
What exactly is medical coding?
Medical coding turns a doctor's description, like chest pain and an electrocardiogram, into standardized numbers that insurers use to decide what to pay. Those code sets are maintained by medical and government groups, and hospitals and practices employ coders or software to apply them to every visit.
What is revenue cycle management?
It is the full chain of billing a health care provider to get paid, from scheduling and insurance checks to coding, submitting claims, and collecting payment. In the U.S., providers often wait weeks or months for insurers to approve or deny claims, so this back office is large and expensive.
Why didn't AI start with diagnosing patients?
Partly caution and partly capability. Diagnosis and treatment carry direct risk to patients and face strict regulation, reimbursement rules, and the need for patient trust. Note writing and claims review were lower risk, highly repetitive, and already digital enough for early AI models to improve.