most genAI initiatives never make it from pilot to production
originally posted to linkedin
most genAI initiatives never make it from pilot to production. true.
therefore genAI isn’t worth the effort and investment. false.
how can this be?
genAI initiatives are asymmetric bets
think of it like a venture capitalist investing in a portfolio of startups. many will fail or break even. The vc invests anyway because a small number of winners can be huge.
the winners tend to get huge because they displace entire workflows.
take stack overflow. in the past when a software developer got stuck, they would post a question to stack overflow. today, they just ask an llm. And stack overflow posts have collapsed from their pre-chatgpt peak. not because devs stopped needing help, but because an entire workflow was displaced.
genAI is still relatively new and immature — both in terms of technology and demand. pilots are built when neither the capabilities nor the use cases are fully understood. failure, iteration, and pivots are not signals of weakness but inevitable rites of passage on the innovation journey. and the only way to arrive at the winners.
so these are the lessons if you are a business leader:
- think in terms of a genAI portfolio, not a single bet
- target workflow displacement, not just augmentation
how does your team measure return on genAI investment?
david crowe - reducibl.com - gatewaystack.com