Insurance is a promise. The screen has to keep it.

Quick Pay is how existing customers pay their premium. More people used it every month than went through the entire buy a new policy journey. And it lost the most people on its very first screen, a form asking for a policy number and date of birth.

My read of the biggest leak was simple. The people leaving at the policy number field were not confused. They just did not have the number with them.

That also explained the losses further down, at identity checks and at payment, where nothing was mechanically broken. People were not only struggling. They were hesitating. You cannot inspect insurance before you buy it, so the screen is the only proof a customer has that the promise is real.

Two designers worked the account. I owned the design system, Quick Pay, the KYC and Bank Update flows, the blog template and eight calculators.

I argued with the client to cut steps out of that flow, because every extra step is a place where someone can leave. The drop off data made the argument, not my opinion. They agreed and the flow was shortened.

Before any of the flows, I built the design system, because four months of payment, verification, blog and calculator work is only possible on a base and there was none.

Take out the blocking field. Mobile number and OTP became the main way in. Policy number stayed as a second option for people who have it.

One decision, not three. The number of screens actually went up, so the honest claim is not fewer steps. Three proceed buttons became one, at the end, with the amount written on it. The popup is gone.

Make the money easy to read. Each installment opens up, is tagged Overdue, Due or Upcoming, and shows the premium, the late fee with GST and the final amount separately. A late fee is the most common reason an amount surprises someone, and hiding it inside a total is how a payment turns into a support call.

The rider step. Nobody asked for this. I suggested it, because riders bring in real money and the payment screen is the one moment a customer is already thinking about their cover. I placed it after the details are confirmed, before the final button, with the cost written plainly and a Skip that is always there.

KYC is three steps when it works. The real work is what happens when it does not.

The Aadhaar OTP, the attempt limit and the two day review are rules from the regulator. I could not remove any of them, so the job was making fixed rules feel clear instead of random. Say what to keep ready before starting. Show attempts left instead of hiding them. Give the waiting state an honest timeline.

The same base carried the blog template, the page type that receives 83% of the site's Google traffic. The old one was plain text with a single call back form. The new one leads with trust, who wrote it and who checked it, then replaces that one form with asks that grow as you read. A calculator early, which costs nothing and shows who is serious. A plan card with Buy Now near the end, once someone has actually read the article.

The numbers above are what our analysts measured before the redesign, not after. The one I would check first is how many people still leave on the first Quick Pay screen, because that is the single number this work was built to move.

The useful version of that lesson is not that I wanted a better team. It is that I found that limit in the middle of delivery instead of at the start.