Opens with a tennis lesson. Coaching a group of four and five year olds, he asked who knew what a forehand was, and a four-year-old turned round and said “it’s this” — then all the others said it too. Not the question he had asked. His point: instruct AI loosely and it makes the same mistake, forgetting things and ignoring things exactly as a person would. From there, twenty-eight minutes with Salesforce’s Sabrina Rodriguez on the argument he keeps returning to — that CMOs are asking what they should use AI for rather than what they are trying to solve, which is the same mistake the industry made with martech a decade ago when it argued about best-of-breed instead of outcomes. On his own team: agents govern themselves, a human governs the governance, and nobody is AI-native yet, including him. He leads differently now because reading the theory and asking the team to apply it turned out not to be a fair ask.
“Agents need to govern themselves in the first instance, but then humans need to govern the governance.”WATCH ON SALESFORCE →