More room for the essence of good project work.
PM leads, AI asks for context.
PM leads, AI finds gaps.
PM leads, AI checks dependencies.
AI drives operations, PM decides.
PM leads review, AI distills learning.
Change requests, new insights, and external impacts flow back into context.
Discussion: Where does friction arise today?
Aligning status between stakeholders — after every meeting and email.
Merging information from meeting, email, and board into one picture.
Reconciling board status with what actually happens in day-to-day project work.
Question: What happens today between meeting and board?
Shared AI project context in the Microsoft environment — a pragmatic starting point for Balluff.
Note: Write access can damage boards quickly — read-only protects the shared truth.
Semantic search across documents and transcripts.
Relationships between epics, dependencies, and roles.
Semantics and structure together — for complex project questions.
LLM-Wiki (Karpathy): growing project memory — each session enriches the context.
This deck is an example: it can be regenerated from the shared project context at any time.
As context grows, the abstraction level of conversations rises.
The value was in speed, clarity, and better follow-up questions.