1 in 10 life insurance policies issued in India are processed by us.

DesiCrew runs new-business policy operations for one of India's largest life insurers. More than 150,000 policies a month are validated and cleared to a single standard, through every seasonal peak, without the client having to manage it. It has run this way since 2007.

A single policy carries a proposal form, a KYC document and a stack of supporting papers, and one mismatch can undo all of it. The tool reads the file — DesiCrew's people decide what goes live. So a family gets paid out on the policy that protects them, and the insurer grows its book without risk building up behind it.

01 — The Mandate

The client is one of India's largest life insurers. Every new policy, whichever channel it comes through, has to clear a single documentation standard before it goes live. DesiCrew runs that operation.

Scale. More than 150,000 policies scrutinized a month — upward of 1.2 million customer documents — and over 800,000 KYC verifications a year. Coverage. Roughly 82% of the client's policy operations, across both offline and online channels. Three lines of work. New-business documents scrutinized against the proposal form, KYC verified, and policy FR processed. Always on. 24 hours a day, every day, sized to whatever the day sends — at an average handling time of 6 minutes per policy scrutiny against a turnaround standard that starts at 2 hours.

02 — The Challenge

The checking is the easy part. The judgment is the hard part, and a family's claim rests on it. Enter a wrong date of birth, or clear a file where the KYC name does not match the form, and the policy still goes live. The break shows up years later, at claim time, when a family is told the payout will not come. One mismatch can undo the whole file.

One policy is not one check. The proposal form is the source of truth; against it sit the KYC and every supporting document, and the name, the address, the date of birth and the identity each have to reconcile, point by point. Most of it is honest error, some is not. Around 90% of what the team catches is a customer or agent filling something wrong, or an agent uploading the wrong person's documents in a batch of a hundred; the rest is the fraud the check exists to stop. The volume will not sit still. Close to 40% of the year's policies land between January and March, then April falls to about a tenth of that.

Anatomy of one policy — why the check cannot be automated

The tool reads the file. Whether a name is close enough, whether a document belongs to this customer or the one before them in the batch, whether a mismatch is a typo or a fraud — these are decisions, not transcriptions. A person makes the call before anything goes live.

03 — The Approach

DesiCrew clears what reconciles, rejects what does not, and builds its own tools to do both faster.

Validate against the form. Every document is checked against the proposal form, the source of truth, before anything moves. Reject, never correct. A file that does not reconcile is rejected and routed to the agent, the bank and the customer at once; DesiCrew validates and hands off to underwriting — it does not fix the file or issue the policy. Flex to the volume. Shifts flex to a load that arrives in real time and cannot be forecast a day ahead, because the team lives close to the centers. Tools built on the floor. Under a contract to raise productivity year on year, the team builds its own tools. Trust, demonstrated. The client hands DesiCrew a volume it cannot predict itself, and signs off on the work at AVP and SVP level.

04 — The Outcome

The claim holds when it is needed. Every policy cleared to standard is one a family can claim on without the payout being contested on a documentation error made at onboarding — so a family's cover is not lost to a mistake made years earlier.

Accuracy held through every peak. The 97% standard is maintained across 1.2 million customer documents a month, including the January-to-March surge — so the book keeps growing without the quality slipping. More cleared by fewer. Industry-best handling times, and a headcount that has come down from 130 to around 100 on comparable annual volume as the tooling compounds — so the cost per policy keeps falling while the standard holds.

05 — Why the relationship holds

DesiCrew began this operation in 2007 with a team of five; it runs on around 106 today. Much of the reliability sits in people who have run this for years — one team lead for a decade. That memory is what lets the team absorb the peak, when it has to grow fast and still not slip.

The client's own side has turned over many times across nineteen years. Companies have come and gone, tenders have opened, an IPO reshaped the vendor list. The work has been re-evaluated again and again and kept every time. It has never been pulled back.

About DesiCrew

DesiCrew is an applied-intelligence company — the human and technology layer that makes AI systems and enterprise operations work reliably and grow at scale, refined in production since 2007. IIT Madras incubated · Everest Group PEAK Matrix 2024 · Great Place to Work.