UX framework and interaction design for scalable human-agent collaboration across multi-agent AI systems.
Design Strategy, UX Framework Design, Interaction Design, AI System Experience
GenAI, LLM, Multi-Agent Systems
Defined UX principles for human-agent collaboration
Designed interaction patterns for transparent AI decision-making
Enabled scalable UX for multi-agent enterprise systems
Defining the Behavioral Principles of Human-Agent Collaboration
Our design team worked with Cisco Outshift to establish a set of behavioral design principles that guide how intelligent agents communicate, collaborate, and build trust with users. These principles form the foundation for designing human-agent workflows that remain transparent, controllable, and reliable.
From Fragmented Interactions to a Scalable Human-Agent Framework
We identified critical gaps in how users understand, trust, and control agent behaviors across workflows. Through a structured design process, we translated these challenges into HAX principles and a reusable component SDK for consistent, scalable experiences.
As AI systems evolve from isolated assistants to networks of autonomous agents, the challenge shifts from model performance to human-agent collaboration design.
Design Strategy, UX Framework Design, Interaction Design, AI System Experience
GenAI, LLM, Multi-Agent Systems
Turning Interaction Principles into a Developer Framework
To help engineering teams implement these interaction patterns consistently, we helped translate the design principles into a structured developer framework. This led to the Agentic Display Interface (ADI) — a schema-based approach that allows agents to communicate through predictable UI components rather than raw text outputs.
Designing the HAX Agent for Multi-Agent Collaboration
As agent ecosystems grow more complex, we explored how interactions could be curated automatically. This led to the concept of the HAX Agent, a behavioral intermediary that filters, prioritizes, and adapts communication between multiple agents and human users.
Expanding How Human-Agent Systems Are Measured
Beyond interface design, the collaboration introduced a new way of evaluating agent-driven systems. Instead of focusing only on model accuracy, the framework emphasizes metrics that measure the quality of human-agent collaboration within real workflows.
.png)
