Human-Agent Centered Design (H-ACD) in Practice
Design practice is undergoing a systemic shift from the production of static artifacts to the governance of behavioral systems where humans and probabilistic agents collaborate over continuous, longitudinal horizons. This paper extends traditional Human-Centered Design (HCD) into Human-Agent Centered Design (H-ACD), establishing a formal framework that operationalizes five primary surfaces: intent, behavior, authority, reasoning surfaces, and longitudinal integrity. By analyzing cross-industry data, including the production-readiness gap in enterprise artificial intelligence, this research synthesizes three composite case studies from enterprise operational command centers, AI-assisted artifact generation pipelines, and agentic software development lifecycles. Each composite combines patterns documented in the cited sources with constructed operating details, so the cases illustrate the framework rather than test it. They yield a practical orchestration framework, ten operating heuristics designed for non-deterministic environments, and a fundamental reframing of the designer's role. H-ACD inherits HCD's core ethical and empathic commitments while critically adapting them to intelligent systems capable of autonomous actuation, escalation, dynamic refusal, and behavioral drift.
Research library