We see value in the idea. Decision traces are crucial because they reveal the observable reasoning behind how decisions were actually made. That said, like much in enterprise AI today, where new breakthroughs seem to emerge every few weeks, we see decision traces as part of the emerging solution to AI decision-making challenges, not as a single magic key. Context graphs only work if they can store enterprise knowledge and map how all organizational data connects.
A part of the picture
The paper identifies a layer we hadn’t paid enough attention to before, however, and that’s important. But we need to broaden the definition to include entities, relationships, provenance, time, permissions, policies—and yes, traces of important decisions, but not only them.
Comparing this to another class of reasoning entities, homo sapiens, helps illustrate the point. Humans rely on different types of memory: episodic memory, which records how decisions were made and what happened; semantic memory, which stores facts and their meanings; and procedural memory, which governs skills and how to perform tasks.
Decision traces fall mostly into the episodic category, but we can’t ignore the other types of reasoning. The semantic layer—the facts and schemas—and the procedural layer—the skills and operating principles—are as important. If we know the facts but don’t understand how decisions were made, for example, it’s hard to reason about future decisions. If we know how decisions were made but not the underlying facts, we can’t ensure conclusions are correct. And if we don’t understand the procedural side, how work is actually done, we’re missing the operational principles people rely on.
