Harmoni Raised $10 Million For Old Factories. The Machines Predate The Internet.
Startup Harmoni, cofounded by David Caputo and Adam Ellis, secured $10 million to bring AI into existing US factories rather than building autonomous ones from scratch. Cal Poly Pomona professor Shokoufeh Mirzaei notes that large companies like Amazon are far ahead on automation, while many small and midsize US factories lack basic analytics and run on legacy systems that predate the internet. Harmoni's bet is that retrofitting intelligence onto existing operations is more practical than greenfield automation.
This demonstrates what I would call the 'brownfield advantage.' The glamorous work in AI draws attention to greenfield deployments, building new systems from nothing. But the actual market opportunity sits in the unglamorous middle: organizations running infrastructure so old it has no analytics layer at all. The lesson is that AI's most profitable applications will not be at the frontier. They will be in the lagging tail, where even basic intelligence is a quantum leap. The frontier is crowded. The tail is not.
Harmoni, cofounded by David Caputo and Adam Ellis, raised $10 million for this retrofit approach. Shokoufeh Mirzaei, professor of industrial and manufacturing engineering at Cal Poly Pomona, provided the assessment that average US factories are behind the automation curve and rely on legacy systems.
- Pick one repetitive task in your own work or home life that you do entirely manually, something you have never thought to automate because it feels too small. A grocery list, a weekly email summary, a budget tracker.
- Open ChatGPT or Claude and describe the task in detail, including the inputs you currently use and the output you need. Ask it to generate a template or repeatable process you can reuse weekly.
- Use that template for two weeks and measure how much time you save. You have just performed a brownfield AI retrofit on your own workflow. The principle is identical to Harmoni's factory play, just at a dramatically smaller scale.