Sukhen Sharma

Scrum Masters and delivery leads spent hours each week pulling Jira data into status reports. By the time leaders read them, the data was stale and reviews were spent arguing over numbers.

I owned the functional design: which sprint, backlog-health, and risk signals the assistant reports, and in what form. I then coached the teams through rollout so the reports replaced the manual ones.

Executive sprint reviews moved from collecting data to making decisions on it. It is the one product in this list with a live user base and a measured outcome.

Street vendors sell all day but rarely know whether they profited. Ledger apps such as KhataBook track what vendors are owed, need typing they have no time for, and are framed around credit they distrust.

I ranked six vendor pains by frequency and severity and scoped v1 to the two a first release could own: profit blindness and capture friction. The vendor speaks the morning purchases once and enters one cash count at night.

Supplier overpay is the biggest money pain, but a price benchmark needs many vendors' data before it means anything. I moved it to v1.1 and built a rule that shows "warming up" instead of a thin, misleading number.

Salon owners run on gut feel because they have no customer history, no station visibility, no inventory control, and no analytics. A competitor teardown of 12+ products showed the same blind spots across the industry.

I wrote a PRD across eight modules, with a Live Station Board as the hero feature. The product works offline with no subscription, because a monthly fee and an internet dependency are what keep first-time users on paper.

Within ten seconds of opening SalonOS each morning, an owner knows what they made yesterday, who is free now, and what is running low. AI features such as a WhatsApp receptionist come in a later phase.

Sellers pay agencies ₹3,000 to ₹15,000 per product and wait three to seven days for images and copy. Listings still get suppressed because sellers do not know Amazon's compliance rules.

Amazon's image and copy rules are enforced on every output as requirements, not left to the seller. The PRD carries three personas and phases from internal testing to a beta with 20 to 50 real sellers.

The PRD lists non-compliant AI output, background-removal failures, and trademarked uploads, each with a mitigation. Pricing model and direct publishing to Seller Central are recorded as open decisions.