IBM’s Watson Health burned $5 billion. Zillow lost half a billion in a quarter. Builder.ai promised automation, then revealed humans were doing the work.
These weren’t failures of technology alone. They were failures of leadership balance.
The pattern is clear: companies keep looking for unicorns. One visionary who can be strategist, technologist, communicator, and evangelist all at once. It doesn’t work. AI is too complex for that mythical figure.
The companies that succeed treat AI leadership as an orchestra, not a solo. They rely on five types of leaders working in concert. Builders are the engineers and data scientists who keep systems running. Strategists see how today’s choices shape tomorrow’s advantage. Translators are product managers and domain experts who help technical and business sides understand each other. Performers are the evangelists who rally support and raise resources. Custodians are risk officers, ethicists, and auditors who protect trust and slow things down when everyone else races ahead.
When one voice dominates, failure ripples outward. Healthcare promised diagnostic breakthroughs but forgot translators and custodians. Tools didn’t fit clinical workflows and patients lost trust. Finance moved at machine speed but forgot custodians, and volatility hit pensions. Education promised personalization but forgot translators, and systems widened inequality instead.
The better model? Walmart. Starting in 2017, they paired builders with domain experts, embedded translators, asked strategists to scale only what worked, and brought custodians in early. The result was hundreds of operational systems that reshaped supply chains and customer service.
The lesson isn’t to avoid naming a Head of AI. It’s to stop looking for unicorns. Orchestrate the five types. Give each a seat. Let none drown out the others.
Read the full piece: https://link.springer.com/article/10.1007/s00146-025-02669-0
