The boring parts are the product

I build financial products in Indonesia, mostly for people the banks ignore. When I describe the work, people expect the interesting parts to be the app, the credit model, maybe the growth loops. They are not. The parts that decide whether the business lives are the license, the audits, and whether someone will trust you with their money or their gold.

Software is cheap to copy now, and AI is making it cheaper by the month. Any feature you ship, a competitor can ship next quarter. What survives is what is structurally hard to get: a scarce license, a control point over the money, a system of record, compliance machinery that took years to build. Which leads to the position I hold more strongly every year: treat regulation as a product surface to win on, not overhead to minimize.

Three things that follow from taking that seriously.

1. Get the hard license before you need it. The lean instinct says launch on the lightest license and upgrade later. In finance that is backwards. The full license is often easier to get before you have customers, and the light one blocks exactly the products, credit and lending, that you will need to actually compete. The months in the regulator's queue feel like lost time. They are the moat forming.

2. Build compliance like a product, not a cost center. If compliance is headcount that scales linearly with customers, the design is unfinished. Give it a roadmap. Turn obligations into software, score risk at the transaction level instead of gating onboarding, automate the evidence. A small team that can prove its compliance paragraph by paragraph can do something the incumbents, with their armies of reviewers, cannot.

3. Refuse the clever balance sheet. The famous finance blowups were rarely product failures. They were duration mismatch dressed up as treasury optimization, borrowing short from customers to hold long. Run the engine as a commission machine that earns on flow, not on float. The yield you give up is the price of durability. Pay it.

None of this demos well. There is no launch video for passing an audit, and no one retweets a license. But the competitive math is hard to argue with: the app took your rival a quarter to copy, and the boring parts would take them years of unglamorous work that most teams will simply refuse to do. That refusal is your protection.

The features are the product for your users. The boring parts are the product for your business. Build both, and be honest about which one is the moat.