Bianca, this is exactly where the AI conversation needs to go next: the principles for splitting work between humans and AI. AI can read, retrieve, summarise, draft, and remind. Humans still carry meaning, judgment, accountability, conflict, and decisions of consequence. Without that distinction, companies either underuse AI or over-automate the wrong things.
The operating model matters as much as the architecture.
The part about knowledge hoarding is the most honest thing in here and the most often ignored in AI rollout plans. The person sitting on decades of tacit expertise has every rational reason not to share it when the incentive structure is competitive, and the fear of being replaced is real. You can't prompt-engineer your way around that. It's a leadership and trust problem first.
The AI consultant will sooner or later reach the point where they coach leaders. This is true, and almost nobody in the AI consulting space is prepared for it. The technical design is often the easier half.
Really glad you started the series here instead of jumping straight into architecture. Looking forward to the next one.
As a company, it is crucial to make it clear that people will not be replaced by AI. That is nothing but hype and total nonsense, because it simply isn't possible. This needs to be understood and clearly communicated. Only then can we consider how to use AI for everyone's benefit. Knowledge must still be shared with everyone, to avoid bottlenecks when only one person has a specific skill, and to ensure everyone keeps developing. Standing still means falling behind.
Bianca, this is exactly where the AI conversation needs to go next: the principles for splitting work between humans and AI. AI can read, retrieve, summarise, draft, and remind. Humans still carry meaning, judgment, accountability, conflict, and decisions of consequence. Without that distinction, companies either underuse AI or over-automate the wrong things.
The operating model matters as much as the architecture.
Yes, and a different mindset means a different architecture and a different architecture means a different operating model.
The part about knowledge hoarding is the most honest thing in here and the most often ignored in AI rollout plans. The person sitting on decades of tacit expertise has every rational reason not to share it when the incentive structure is competitive, and the fear of being replaced is real. You can't prompt-engineer your way around that. It's a leadership and trust problem first.
The AI consultant will sooner or later reach the point where they coach leaders. This is true, and almost nobody in the AI consulting space is prepared for it. The technical design is often the easier half.
Really glad you started the series here instead of jumping straight into architecture. Looking forward to the next one.
As a company, it is crucial to make it clear that people will not be replaced by AI. That is nothing but hype and total nonsense, because it simply isn't possible. This needs to be understood and clearly communicated. Only then can we consider how to use AI for everyone's benefit. Knowledge must still be shared with everyone, to avoid bottlenecks when only one person has a specific skill, and to ensure everyone keeps developing. Standing still means falling behind.