"The real issue is it that it takes time to implement changes in processes, skills and organizational structure to fully harness AI’s potential"

Not to mention a core issue- legacy systems, data silos and low-interoperability need to be addressed before AI can address big enterprise issues. Until then it is best applicable for low hanging fruit projects. See "AI- don't be seduced by the hype!"

AI has not had an impact on overall productivity yet ( though it will in the longer term) whilst augmenting human intelligence is a faster route to efficiency.

"As other researchers and managers have also observed, we are finding that most machine learning applications augment, rather than replace, human efforts. In doing so, they demand changes in what people are doing. And in the case of AI — even more than was true with ERP systems — those changes eliminate many nonspecialized tasks and create skilled tasks that require good judgement and domain expertise."

See "The human flaw in AI implementation"

Just as with other "hot" technology there is no panacea nor any alternative than to fit AI as a means to end end in parallel with commercial, organisational, structural and innovation strategies.