What we have learned shipping this stuff.
Written by the people doing the work, about the parts that decide whether a build survives production. No thought leadership, no predictions.
Most agent projects fail in the same place: nobody decided what the agent is allowed to do on its own. That is an engineering decision, not a prompting one.
Most businesses already have an integration layer. It is a human being, and they are the least reliable and most expensive part of the stack.
Large transformation programmes fail quietly for eighteen months. A small slice in production fails loudly in three weeks, which is the useful kind.
The second question every operator asks is what it did and why. If the system cannot answer that, the demo does not matter.
If a build has no agreed metric, its success becomes a matter of opinion. Opinion always favours whoever is presenting.
A folder of contracts is not something you can ask questions of. Getting from one to the other is data engineering.