Adoption pressure is real. Boards ask about artificial intelligence, competitors announce initiatives, and vendors present capability that is genuinely impressive. The risk is that the organisation selects a technology before it has articulated the problem the technology is meant to address.
When that happens, the implementation usually survives as a parallel activity. It exists alongside the operation rather than inside it, requires its own maintenance, and produces outputs that no existing process depends on. It is not a failure of the technology; it is a failure of sequencing.
The alternative sequence is unglamorous but reliable. Establish where the organisation loses time, capacity or accuracy. Determine which of those problems are structural and which are informational. Apply automation to the structural ones and artificial intelligence to the informational ones. Measure against the original problem rather than against adoption.
This also protects the organisation commercially. A transformation priority survives changes in technology; a technology commitment made without a priority does not.