I think the consensus here is pretty telling... Diagnosis, judgment, prioritization, system design, constraints, measurable outcomes. Notice what almost nobody said? "Building." That may be the real shift. As implementation gets cheaper, the builder is no longer the scarce resource. The scarce resource becomes the person who can walk into a messy business, figure out what is actually broken, separate symptoms from causes, decide what is worth fixing, design the right solution... and sometimes tell the client, "don't build anything." And I think there's one more layer beyond diagnosis... Ownership. It's easy to recommend an automation. It's harder to put your name behind the outcome. Did it save money? Did it make money? Did it reduce risk? Did it remove a constraint? Did the people actually use it? Was it worth doing in the first place? AI can generate 20 possible solutions before lunch. The valuable person may increasingly be the one who knows which 19 to throw away. That skill isn't really "AI consulting." It's business judgment... with AI as one hell of a toolbox.