Product Manager + Engineer + AI = Builder
SmaranChallapalli
Every complex product problem is just 10–15 smaller problems pretending to be one.
Co-running Fisique on the floor and BotBrained on the screen. Previously Paytm, the login screen for 100M+ every month.
scroll ↓
Work
12 entriesHow I work
Operator first, PM second.
Most mornings start on the gym floor in Kokapet at 6 a.m., front desk, fingerprint reader, the member who didn't show. The product decisions that hold up are the ones written by someone who'll have to live with them on Monday.
Build the thing before the doc about the thing.
A working prototype is a better spec than a spec. The PRD reads honest when it describes something real instead of something hoped for.
Write down what good looks like, before anyone codes.
If I can't put success in a line someone else can verify, it isn't ready to build. The bar gets agreed on first, in writing, in front of the people who'll be measured against it.
The unhappy path is the product.
At Paytm scale you learn this the hard way: the SMS fails, the device unbinds, the bank rail times out, the regulator calls. The boring resilience work is most of the work. Happy paths sell demos; unhappy paths keep the lights on.
Self-serve, or it's a service not a product.
If a new user can't reach their first win without me on a call, I haven't built a product yet. I design for the lonely Tuesday-night signup who'll never speak to anyone on my side.
AI is a teammate, not autocomplete.
I work alongside Claude and Cursor every day, delegating, reviewing, pushing back. This site was wireframed and shipped that way. So was most of LeadVoice. The leverage isn't typing faster; it's running three threads in parallel and keeping all three honest.
Keep the shelf in public.
VirOrb got shut down when the unit economics broke. UPICred got parked after the regulatory cost became the wrong fight. Both are on this page on purpose. Pretending the shelf is empty is worse than admitting it has shelves. The scars are the part that compounds.
Distribution beats novelty.
A clever product nobody can find is a hobby. Kokapet.run became Fisique's best lead source not because the idea was new but because Sunday mornings are loud and Instagram is free.
None of this is a methodology. It's the residue of nine years of shipping things that worked, and a few that didn't.