HBR On Strategy The Right Way to Launch an AI Initiative
66 snips
May 7, 2025 Iavor Bojinov, an assistant professor at Harvard Business School and former LinkedIn data scientist, delves into the challenges of launching AI initiatives, which often fail due to their complexity. He outlines five key steps for success: selection, development, evaluation, adoption, and management. The discussion highlights the importance of trust between users and AI, balancing speed with ethical standards, and the critical role of experimentation. Case studies from Etsy and LinkedIn reveal valuable lessons about user interactions and strategy alignment.
AI Snips
Chapters
Transcript
Episode notes
AI Project Challenges
- AI projects differ critically from traditional IT projects due to their probabilistic, non-deterministic nature.
- This uncertainty causes higher failure rates and multiple failure points unique to AI, such as poor accuracy and biased outcomes.
AI Adoption Failure at LinkedIn
- Iavor Bojinov built an AI tool at LinkedIn that cut analysis time dramatically but users refused to adopt it.
- The key barrier was lack of trust despite clear utility, illustrating "if you build it, they will not come."
Prioritize Impact and Ethics
- Start AI projects by prioritizing impact over feasibility; ensure strategic alignment with business goals.
- Address ethical considerations like privacy, fairness, and transparency upfront to avoid costly rework.