When Would Appreciative Inquiry not Work? And Why?

Circumstances in which Appreciative Inquiry Won't Work
JC de Jong, SIG Leader
AI does work and many, many interventions have proven that. See amongst others the cases in the book of Watkins, Mohr and Kelly "Appreciative Inquiry".
It feels that you very much answer your own question as to why change interventions don't work. When there is no earnest and open participation possible all change initiatives will in the long(er) term fail and that goes for AI as well.
We can also say that in such cases AI has not been applied as it is intended to. When you look at the 5 AI principles (see the summary in the AI knowledge center) then it becomes clear that it requires an open and earnest participation of all people (the community, including management) invited to participate in the process.
So far after nearly 20 years of applying AI I can boldly state that whenever this open attitude was applied AI worked.

Appreciative Inquiry is Just Another Tool in the Toolkit
Gary Wong, Premium Member
Here are a couple of circumstances when you should not use AI:
1. When you as the facilitator must not contaminate the data. Because we move in the direction we inquire, by asking for positive stories we influence the conversations. As an AI practitioner we must be on guard that because we are good at using a hammer, we see everything as a nail. If not careful, our AI expertise can become a cognitive bias.
2. When we want to learn from failures. There are arguments that we actually learn more from our failings than our successes.

I now am realizing that Appreciative Inquiry is a subset of Narrative Inquiry. In NI, we act as story prompters and are willing to gather both positive and negative stories. We should be empowering the storytellers to move in the direction they, not us, wish to go. If the morale is low and all you hear are negative stories, then prompt for positive success stories. On the other hand, if it's all about feeling good, I'll ask for stories of failure.