How do you sell a prototype to a bank or large institution?

The reason the cycle is so long is what's at stake: customers are handing over "the keys to their hard earned projects" — assets where infrastructure investors have spent a hundred to two hundred million dollars per transaction expecting a minimum 12% internal rate of return. It takes a long time for somebody to build that trust in you.

Rakhesh's playbook is track record plus radical honesty. In energy, after you leave the buyer's office they get on LinkedIn, find shared connections, and call to ask "what do you know about this guy, is he any good?" His edge: "I've never lied, I've always been me, I've always been genuine and I've always delivered" — and when he failed at something, he was open about why. "Anyone that's out there saying they've got all the solutions is lying. But if you can show to a customer that you've lived it, and that's where your solution comes from, then you have a much greater chance of success."

Satya contrasts the approaches: with another startup "you just go knock their door" and ask if they'll use it, but a public company or a bank is quite different. Since Socratix AI knew it would be working with larger institutions — fintechs, banks, credit unions and marketplaces — they decided to build the product "with security and long-term scalability in mind" and started getting compliance done from the very first couple of weeks. On deal structure — paid pilots, design partnerships, testing on production data or backtesting on historical data — his rule is to do "whatever gets your customer to use your product." If you're confident enough in the problem and the product, they'll keep buying, so you should optimize for the long term, not the short term.