sandeep.khanna
search everything⌘K
/thinking

Thinking

Notes that explain how I think. Not polished — working files.

/thinking/healthcare/Insurance.md

Insurance

Insurance claim automation looks like a data problem. It’s actually a trust problem between three parties who each have a reason not to be the one who made the mistake.

I spent time going deep on India’s NHCX infrastructure, specifically whether a policyholder could get a formal, timestamped, auditable record of notifying their insurer of hospitalization for a non-network hospital. The honest finding: the BSP (Beneficiary Service Platform) role exists in the HCX Protocol spec, but it isn’t activated yet in NHA’s live production network. The clean, direct version of this product doesn’t have a v1 path today.

What does have a v1 path is a hospital-assisted TSP model, routing the notification through the hospital’s existing systems rather than building a new consumer-facing rail from scratch. Slower to build the ideal version, faster to build something that actually works with the infrastructure as it exists, not as the spec says it will eventually exist.

The part that reframes the whole idea: insurers already can’t reject claims purely on intimation-delay grounds. There’s Supreme Court precedent and an IRDAI circular that already limit this. Which means the real value of a “proof of notification” product isn’t closing a legal loophole. Most of that loophole is already closed.

It’s removing the friction and anxiety of not knowing whether you did the paperwork right, at the exact moment a family has zero bandwidth to care about paperwork.