Agentic Social Affordance FrameworkASAF
A theoretical framework by Meng-Han Lee (Zaious)
“Agent identity design is not a UX convention.
It is a collaboration interface.”
Overview
The Agentic Social Affordance Framework (ASAF) proposes that in multi-agent AI systems, the social identity of individual agents functions as a structural collaboration interface — shaping how human operators perceive, approach, and engage with each agent.
When systems scale beyond human working memory capacity (approximately 4–7 agents, depending on chunking strategy and individual differences), social identity design shifts from a superficial preference to a structural necessity — an indispensable design consideration rather than a guaranteed effect, with magnitude moderated by individual cognitive style (anthropomorphizing vs. instrumentalizing).
ASAF treats the separability of the social affordance layer and the engineering orchestration layer as a framing assumption — an organizing distinction for design analysis, not a testable claim of effect-independence: the two layers' downstream effects can and do interact. And where classical multi-agent systems used roles, norms, and coordination to constrain autonomous agents, ASAF reverses the direction — the same organizational vocabulary structures the cognition and oversight of the human operators who remain in the loop.
Three Mechanismscore idea
1. Identity Signaling
Pre-interaction role activation through codename, archetype, and characteristic utterances. Before a user sends a single message, the agent's social identity has already shaped expectations, vocabulary, and interaction strategy.
2. Behavioral Priming
Sustained improvement in user input quality driven by social role expectations. Users provide more precise, domain-relevant instructions when they perceive the agent as a specialist rather than a generic tool.
3. Collaborative Governance
Differential oversight calibration based on agent topology. Users naturally scrutinize auditors more than ideators, creating an emergent quality assurance layer without explicit rules.
Identity Signal Fidelity Spectrum
The four tiers are illustrative — sampled points along a continuous spectrum, not a developmental trajectory. Deployment tier reflects operational requirements, not technological maturity. ASAF's strongest predictions apply at Tiers 3–4; they apply partially at Tier 2, and Tier 1 is a boundary condition where effectiveness degrades with identity-signal fidelity.
Case Study: ChronicleCore
ASAF emerged from hands-on experience building ChronicleCore, a production multi-agent system with 38 specialized AI agents governed by a 5-Pillar architecture. The system operates at Tier 3 (Structured Identity Enforcement), where each agent maintains a persistent social identity through codename, archetype, personality traits, and domain-specific expertise.
The design illustration — The Inquisitor (真理), ChronicleCore's sole adversarial node — shows how an explicitly non-cooperative social identity is used to structure operator oversight. As a Hypothesis & Theory contribution, ASAF formalizes its claims as testable hypotheses; empirical validation is outlined as future work, not claimed.
Publication Status
How to Cite
Author
Meng-Han Lee (Zaious) — AI Agent Architect & Independent HCI Researcher, Taipei, Taiwan.