Description
Decomposes complex user behaviour into many simple interacting 'agents' (sub-processes with goals), revealing conflicts and coordination failures that explain behaviour.
Minsky's Society of Mind proposed that intelligence arises from the interaction of many simple, mindless agents, each performing a narrow function, organised into agencies with conflicting and cooperating goals. As an analytic method, agent decomposition models a user's behaviour in a complex situation as a set of such agents (for example, a goal to save, a habit to browse, a need for social approval) and maps how they compete, which wins in which context, and what arbitrates between them. This makes inconsistent behaviour explicable and suggests interventions that strengthen, weaken or coordinate specific agents. Rigour comes from grounding each agent in observed evidence, specifying activation conditions, and testing predictions against behaviour.
Attributes
- Discipline
- Task Analysis & Cognitive Modelling
- Loss type
- Extinct
- Stage
- Definition
- Data
- Analytical / expert
- Effort
- Days
Steps
- Agent-level error analysis: for any UX error, decompose into agent-level failures:
- Attention agent: was the relevant information in the attentional field?
- (2) Recognition agent: was the stimulus correctly identified?
- (3) Memory agent: was the correct action retrieved?
- (4) Execution agent: was the action performed correctly?
- (5) Goal agent: was the current action aligned with the active goal? Each agent failure type maps to a specific design intervention.
Sources
- The Society of Mind — Marvin Minsky
Also known as
- Society of Mind analysis
- Agent-based problem decomposition
Categories
- User Research
- Task Analysis & Cognitive Modelling