You Cannot Predict the Head Pin.

People ask me which segment will adopt agentic commerce first. I do not know yet. Nobody does. You cannot predict the head pin. You uncover it. I learned that at my first company. Vivox taught me what to build while you wait.

Geoffrey Moore’s bowling alley works like this. Win one narrow segment of pragmatists with a painful problem, then use it to knock over the segments behind it. The first segment is the head pin.

In 1994 NETworth’s users were mostly other students and a handful of curious early adopters. Then the Mosaic team at NCSA called. They wanted NETworth in the browser’s drop-down menu. I approved it on the spot, and when Microsoft licensed the Mosaic code for Internet Explorer, the menu came with it.

In 1995 I pitched the same thing to two of the most powerful firms in finance. At Merrill Lynch the chief executive slapped the table and called it a fad. Two weeks later, at Charles Schwab, the room leaned in and signed a contract. Same pitch. Same product. By the time Intuit called, about a hundred financial institutions used NETworth, including Schwab, Vanguard and 20th Century. That is what falling pins look like.

I did not plan either of those moments. The browser team found us. Schwab leaned in while Merrill leaned back. Even the smartest room I ever sat in, at the Agenda conference in 1996, could not see which search engine would win. I tell that story in Nobody Raised a Hand.

Clayton Christensen drew the practical conclusion in The Innovator’s Dilemma. Plans for disruptive change should be plans for learning and discovery, not plans for execution.

If you cannot predict the head pin, you have to be ready for whichever one falls. Vivox taught me that by losing. I was chief product officer. We built a powerful voice platform for games. It did spatial audio, conference calling and two-factor authentication. It was also monolithic. You took the whole system or nothing. Twilio broke many of the same functions into small blocks any developer could pick up one at a time. Vivox was sold to Unity, a base hit, maybe a double. Twilio became a multibillion-dollar company. The difference was not the technology. It was the shape.

So trust for agents gets built as blocks: an identity block, a facet block, a policy block, a payment proof block. Each one usable in an afternoon. Our job is not to guess right. Our job is to build the blocks so that whoever finds the head pin, and whatever it turns out to be, can pick them up and use them. I lost that round once. I do not intend to lose it twice.