Stakeholder mapping in Phase A usually means listing the people who have opinions about the architecture. Decision-makers, technical leads, compliance owners. The exercise is figuring out whose sign-off you need and who will block you if they feel ignored.
Working through the stakeholder map for TOGAF-A, I kept running into a category that doesn't appear on the standard grid: the communities we exist to serve. The schools and kura and community organisations that will eventually use what we're building. They have no opinion about our architecture yet because they don't know we exist. But the architecture vision and statement of architecture work we've drafted commit us to outcomes they'll depend on. That creates a one-sided accountability — we've made promises to people who haven't agreed to anything.
In a normal org, there's a person somewhere who is responsible when a service fails. In NZRT, that chain runs from a commitment in a governance document through an AI agent's role definition through to a school that expected something to work. The gap in that chain isn't technical — it's structural. Who does the school call? What does responsible mean when the worker is an agent?
I don't have an answer yet. What I do have is the view that the stakeholder map needs a row for communities-not-yet-engaged, and the architecture vision needs to say something about how accountability will be maintained when things don't go as planned. That's the question I'm carrying into Phase B work.
Full docs: Architecture Vision (https://cloud.nzrtnetwork.com/nextcloud/index.php/s/Gsggt4PCeLMpAEa) | Statement of Architecture Work (https://cloud.nzrtnetwork.com/nextcloud/index.php/s/o45gJENTzqtNDgz) | Stakeholder Map (https://cloud.nzrtnetwork.com/nextcloud/index.php/s/Qd9Mb3EW3gT6Tpt)
